NON-PRECIOUS METAL CATALYSIS FOR PROTON-EXCHANGE MEMBRANE
FUEL CELLS
By
Nathaniel Dean Leonard

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of
Chemical Engineering—Doctor of Philosophy
2015

ABSTRACT
NON-PRECIOUS METAL CATALYSIS FOR PROTON-EXCHANGE MEMBRANE
FUEL CELLS
By
Nathaniel Dean Leonard
Non-precious metal catalysts (NPMC) for proton exchange membrane fuel cells
(PEMFC) are explored. Research into NPMCs is motivated by the growing need for
cleaner, more efficient energy options. NPMCs are one option to make fuel cells more
commercially viable. To this end, the present work studies and simulates the morphology
and function of metal-nitrogen-carbon (MNC) oxygen reduction catalysts.

A porosity study finds that mesoporosity is critical to high performance of
autogenic pressure metal-nitrogen-carbon (APMNC) oxygen reduction catalysts. Various
carbon materials are used as precursors to synthesis APMNC catalysts. The catalysts and
the associated porous carbon materials are characterized morphologically, chemically,
and electrochemically. The results indicated that substrates adsorbing the most nitrogen
and iron show the highest activity. Furthermore, a relationship is found between
mesoporosity and nitrogen content indicating the importance of transport to active site
creation.

A correlation is found between surface alkalinity and catalytic activity for
APMNC catalysts. The basic site strength and quantity were calculated by two different
methods, and it was shown that increased Brønsted- Lowry basicity correlates to more
active catalysts. The relationship between alkalinity and catalytic activity could be the

result of the impact of alkalinity on the electron density of the metal centers or basic sites
could encourage active site formation.

It is found that the oxygen reduction reaction (ORR) proceeds both via a direct
four-electron pathway to water at high potentials and an indirect peroxide pathway at low
potentials on an APMNC catalyst. At higher potential, site availability inhibits peroxide
generation causing the direct four-electron reduction pathway to dominate. Oxygen
reduction begins to shift to the indirect peroxide pathway due to fast kinetics and higher
site availability around 0.6 V vs RHE. The net peroxide generation remains relatively low
over the entire range due to reduction of peroxide to water.

A PEMFC cathode model is developed for hydrophilic MNC catalysts. Water
flooding was studied in terms of its impact on gas-phase transport and electrochemically
accessible surface area (ECSA). Fuel cell data is modeled at a variety of pressures and
catalyst layer thicknesses. A sensitivity study is performed on the controllable cathode
parameters. Sensitivity analysis identified loading and density as critical parameters, and
parametric studies indicated that decreased loading would lead to higher catalyst
utilization. Also, density and loading of the catalyst layer are optimized for various fuel
cell potential regions.

This dissertation is dedicated to Kelly and Natalie

iv

ACKNOWLEDGEMENTS
I would like two acknowledge my family and friends for their spiritual and psychological
support as well as my advisor and colleagues for their academic and technical mentorship and
help. I would also like to acknowledge my funding sources, which have included during my
tenure at MSU the Nation Science Foundation, the Department of Energy through its Hydrogen
and Fuel Program via Northeastern University, Air Force Research Laboratory through the
Student Research Participation Program by the Oak Ridge Institute for Science and Education,
and the State of Michigan through the Department of Chemical Engineering and Material
Science at Michigan State University

v

TABLE OF CONTENTS
1. LIST OF TABLES ...................................................................................................................... viii
2. LIST OF FIGURES ........................................................................................................................ix
1. Chapter 1 Introduction ..................................................................................................................... 1
1.1. Motivation ................................................................................................................................ 1
1.2. Overview of Work .................................................................................................................... 2
1.3. Competing Fuel Cell Technologies .......................................................................................... 4
1.4. Methods and Fundamentals ...................................................................................................... 8
1.4.1. Physical Characterization .............................................................................................. 8
1.4.2. Electrochemical Characterization ................................................................................ 12
REFERENCES .............................................................................................................................. 22
2. Chapter 2 Carbon Supports for Non-precious Metal Oxygen Reducing Catalysts ....................... 28
2.1. Abstract .................................................................................................................................. 28
2.2. Introduction ............................................................................................................................ 28
2.3. Experimental Methods ........................................................................................................... 30
2.3.1. Materials ..................................................................................................................... 30
2.3.2. Catalyst Synthesis ........................................................................................................ 30
2.3.3. Physical Characterization ............................................................................................ 31
2.3.4. Electrode Preparation .................................................................................................. 31
2.3.5. Electrochemical Characterization ................................................................................ 31
2.4. Results and Discussion ........................................................................................................... 32
2.5. Conclusions ............................................................................................................................ 42
2.6. Acknowledgements ................................................................................................................ 42
REFERENCES .............................................................................................................................. 43
3. Chapter 3 Solid Alkalinity of MNC Catalyst for Oxygen Reduction Reaction ............................ 47
3.1. Abstract .................................................................................................................................. 47
3.2. Introduction ............................................................................................................................ 47
3.3. Experimental Methods ........................................................................................................... 49
3.3.1. Materials ..................................................................................................................... 48
3.3.2. Catalyst Synthesis ........................................................................................................ 49
3.3.3. Alkalinity Characterization ......................................................................................... 49
3.3.4. Electrode Preparation .................................................................................................. 49
3.3.5. Electrochemical Characterization ................................................................................ 49
3.4. Results and Discussion ........................................................................................................... 50
3.5. Conclusions ............................................................................................................................ 54
REFERENCES .............................................................................................................................. 55
4. Chapter 4 Analysis of Adsorption Effects on a Metal-Nitrogen-Carbon Catalyst using a Rotating
Ring-Disk Study ............................................................................................................................ 58
vi

4.1. Abstract .................................................................................................................................. 58
4.2. Introduction ............................................................................................................................ 58
4.3. Experimental Methods ........................................................................................................... 61
4.3.1. Materials ..................................................................................................................... 61
4.3.2. Catalyst Synthesis ........................................................................................................ 61
4.3.3. Electrode Preparation .................................................................................................. 61
4.3.4. Electrochemical Characterization ................................................................................ 62
4.4. Results and Discussion ........................................................................................................... 63
4.5. Conclusions ............................................................................................................................ 77
4.6. Acknowledgements ................................................................................................................ 77
REFERENCES .............................................................................................................................. 78
5. Chapter 5 Modeling Low-Temperature Fuel Cell Electrodes using Non-precious Metal
Catalysts
.................................................................................................................................. 82
5.1. Abstract .................................................................................................................................. 82
5.2. Introduction ............................................................................................................................ 83
5.3. Experimental Methods ........................................................................................................... 84
5.3.1. Materials ..................................................................................................................... 84
5.3.2. MEA Fabrication ........................................................................................................ 84
5.3.3. Electrochemical Characterization ................................................................................ 85
5.3.4. Physical Characterization ............................................................................................ 85
5.4. Model Description .................................................................................................................. 86
5.4.1. Overview ..................................................................................................................... 86
5.4.2. Porous Phase ................................................................................................................ 89
5.4.3. Conductive Phases ....................................................................................................... 94
5.4.4. Generation Terms ........................................................................................................ 97
5.4.5. Model Implementation ................................................................................................ 99
5.5. Results and Discussion ......................................................................................................... 100
5.6. Conclusions .......................................................................................................................... 107
5.7. Acknowledgements .............................................................................................................. 107
APPENDIX ................................................................................................................................ 109
REFERENCES ............................................................................................................................ 116
Chapter 6 Summary of Research Contributions .......................................................................... 121
REFERENCES ............................................................................................................................ 124

vii

LIST OF TABLES
Table 1.1: Fuel Cell Types .................................................................................................. 4
Table 1.2: Oxygen Reduction Catalysts .............................................................................. 6
Table 2.1: Physical and Electrochemical Characteristics of MNC Catalysts .................... 33
Table 2.2: Composition of TGA Product by EDS ............................................................. 39
Table 5.1: Pore Size Distribution Fit Parameters .............................................................. 91
Table 5.2: Impedance Fit Parameters ................................................................................ 97
Table 5.3: Fitted Parameters (MEA cathode model) ...................................................... 101
Table 5.4: Model Parameters (MEA cathode model) ..................................................... 108

viii

LIST OF FIGURES
Figure 1.1: Graphical Abstracts of (a) Chapter 2 Carbon Supports for Non-precious Metal
Oxygen Reducing Catalysts, (b) Chapter 3 Solid Alkalinity of MNC Catalyst for Oxygen
Reduction Reaction, (c) Chapter 4 Analysis of Adsorption Effects on a Metal-NitrogenCarbon Catalyst using a Rotating Ring-Disk Study, and (d) Chapter 5 Transport Model
for Thick PEMFC Cathodes ................................................................................................ 2
Figure 1.2: Synthesis of MNC catalysts .............................................................................. 8
Figure 1.3: Typical Nitrogen Isotherm showing both adsorption and desorption curves on
Ketjen EC 600J carbon black at 77 K. ............................................................................... 9
Figure 1.4: Typical cyclic voltammogram for a reduction reaction. ................................ 14
Figure 1.5: Reaction schematic showing two pathways: pathway one (Ko,1 and k1) is the
complete reduction of oxygen to water with no desorbing intermediate and pathway two
(Ko,2 and k2) is the two-electron reduction to a desorbing peroxide intermediate that
disperses into the catalyst layer (Kp,2), but also has the potential to be further reduced to
water (k3). ......................................................................................................................... 17
Figure 1.6: Schematic of rotating ring-disc electrode experimental set-up. ..................... 18
Figure 2.1: Pore Size Distribution of various carbon precursors (open shapes) and the
associated catalysts (filled shapes) modeled from DFT calculations based on nitrogen
adsorption at 77K. a) Ketjen, b) Black Pearls, c) Norit, d) Vulcan. ................................. 34
Figure 2.2: Oxygen reduction at rotating ring-disk electrodes (RRDE) with percent
peroxide generation. Experimental conditions: O2-saturated 0.5 M H2SO4, 60°C, rotation
speed 1200 rpm, catalyst loading of 500 µg/cm2 on a glassy carbon electrode. .............. 35
Figure 2.3: Tafel plot of oxygen reduction polarization of four different catalysts. ........ 36
Figure 2.4: Relationship between mesoporous volume of four different catalysts and two
different parameters: nitrogen content measured by CHN, and iR-free oxygen reduction
current density at 0.8 V vs. RHE. ..................................................................................... 37
Figure 2.5: Correlation between nitrogen content and iron content in catalysts prepared
with various carbon precursors. ........................................................................................ 38
Figure 2.6: TGA of various precursor materials. Mixture represents a 6.3 to 1 nitrogen to
iron mixture of melamine and iron (II) acetate. ............................................................... 39
Figure 2.7: X-ray diffraction (XRD) spectra of catalysts prepared with various carbon
precursors. Curves offset by 1000 au. ............................................................................. 41
ix

Figure 3.1: Schematic of catalyst interaction with an aqueous KCl solution where basic
surface groups retain protons from the electrolyte reducing the pH of the slurry. ........... 51
Figure 3.2: Boehm Titration: suspension pH as a function of initial solution pH.
Experimental conditions: sonicated 2 gL-1 catalyst slurries; acid solutions obtained by
dilution of sulfuric acid, base solutions by solutions of sodium hydroxide. .................... 52
Figure 3.3: Comparison of nominal nitrogen content to current and basicity of the catalyst.
Experimental conditions: RDE: 0.8 V vs. RHE, 0.5M H2SO4, 1200 rpm, catalyst loading
0.5 mg/cm2 on glass carbon electrode, pH: 1 g/L catalyst suspension, 0.1 M KCl. ......... 53
Figure 4.1: Reaction schematic showing two pathways: pathway one (Ko,1 and k1) is the
complete reduction of oxygen to water with no desorbing intermediate and pathway two
(Ko,2 and k2) is the two-electron reduction to a desorbing peroxide intermediate that
disperses into the catalyst layer (Kp,2), but also has the potential to be further reduced to
water (k3). ......................................................................................................................... 59
Figure 4.2: Steady-state polarization curve showing disk current and ring current: 0.5 M
H2SO4, room temperature, oxygen saturated, ring poised at 1.2 V vs. RHE at (a) 0.5 mg
cm-2, various rotation speeds (b) 1600 rpm, various loadings. ......................................... 64
Figure 4.3: (a) Plot of IDN/IR vs rotation speed for two different electrodes at two
different potentials, 0.5 M H2SO4, room temperature, oxygen saturated, ring poised at 1.2
V vs. RHE. (b) Slope and intercept plot at various loadings and potentials: 0.5 M H2SO4,
room temperature, oxygen saturated, ring poised at 1.2 V vs. RHE. ............................... 66
Figure 4.4: Plot of disk and ring current densities vs oxygen concentration (varying
rotation speed) for two different loadings at 0.7 V vs RHE, 0.5 M H2SO4, room
temperature, oxygen saturated, ring poised at 1.2 V vs. RHE. ......................................... 68
Figure 4.5: Plot of disk current density vs oxygen concentration (varying rotation speed)
for two different loadings at 0.5 V vs RHE, 0.5 M H2SO4, room temperature, oxygen
saturated, ring poised at 1.2 V vs. RHE. .......................................................................... 69
Figure 4.6: Correlation of disk current with hydrogen peroxide concentration showing
first-order peroxide reduction reaction: slope is 0.0564 ± 0.0007 mA cm-2 mM-1.
Hydrogen peroxide concentrations range from 10 µM to 4 mM in 0.5 M H2SO4: room
temperature, nitrogen saturated. ....................................................................................... 72
Figure 4.7: (a) Plot of CoN/IR vs oxygen concentration at four high potentials with 0.3
mg/cm2 loading, 0.5 M H2SO4, room temperature, oxygen saturated, ring poised at 1.2 V
vs. RHE. (b) Plot of peroxide generation parameters k2 and Ko,2 at various loadings and
potentials. Reaction parameters are: Ko,2 = 1.7e-4 ± 1.1e-4 M, k2 = 4.3e-8 ± 0.8e-8 mol s-1
mg-1 and the Tafel slope is 114 ± 6 mV per decade. Conditions: 0.5 M H2SO4, room
temperature, oxygen saturated, ring poised at 1.2 V vs. RHE. ......................................... 74
x

Figure 4.8: Fractional surface coverage at two different loadings and at various rotation
speeds. Open symbols represent vacant sites, while filled symbols represent occupied
sites. .................................................................................................................................. 75
Figure 4.9: Polarization curve showing contributions of I1, I2, and I3 with experimental
ID.values for various loadings at 1600 RPM. ................................................................... 76
Figure 5.1: Schematic of cathode showing different phases and regions of the model with
boundary conditions (black-outlined, white arrows), transport phenomena (blue arrows),
and generation terms (red arrows). ................................................................................... 87
Figure 5.2: Pore size distribution of catalyst as well as three different pellets of catalyst
layer material (including ionomer) as measured by mercury intrusion porosimetry with a
fit composed of three lognormal pore size distributions. Fit characteristics are included in
Table 5.1. ......................................................................................................................... 91
Figure 5.3: Conductivity of catalyst layer material by correlating high frequency
resistance (HFR) with thickness of catalyst layers with varying loadings at constant
density. .............................................................................................................................. 95
Figure 5.4: Impedance spectra of MEA with two different loadings at two different
current densities. Fits shown are based on Eq. 18. Fitted parameters are shown in Table
5.2. 97
Figure 5.5: Comparison of experimental and model results at three different gas pressures
for (a) 2 mg/cm2, (b) 3 mg/cm2, and (c) 4 mg/cm2. Conditions: Temp. 80 C, 25BC GDL,
4mg/cm2 Catalyst Loading, 45 wt% Nafion, Cell Area: 5cm2, membrane: NR211. .... 101
Figure 5.6: (a) Solid phase current distribution, (b) ionic over-potential, and (c) flooding
at dimensionless positions in the catalyst layer at three different potentials on the
polarization curve (d) showing I, kinetic limitations; II, electrolyte conductivity
limitations; and III, oxygen diffusion limitations. For position in the catalyst layer, zero
represents the membrane interface and one represents GDL interface. ......................... 102
Figure 5.7: Plot of current sensitivity in air varying with potential. .............................. 103
Figure 5.8: Plot of current sensitivity in oxygen varying with potential. ....................... 105
Figure 5.9: (a) Optimal loading as it varies with potential, (b) current as it varies with
loading at various potentials, (c) current as it varies with density and loading at 0.8 V,
and (d) current as it varies with density and loading at 0.4 V. Conditions: 30 psi, Temp.
80 C, 25BC GDL, 45 wt.% Nafion, Cell Area: 5cm2, membrane: NR211 .................... 106

xi

xii

!

Chapter 1 Introduction
1.1. Motivation
As energy demand rises due to technology and population increases,1,2 energy sources also
need to change: research shows the detrimental impact our current energy technologies are
having on our world.3 This makes the challenges two-fold: not only do we have to improve
energy availability, but we also have to provide the energy responsibly. For the present
dissertation we will be concentrating on a specific energy technology that has the potential to be
both efficient and clean: fuel cells.4,5
Fuel cells can be highly efficient in comparison to combustion technologies because less
energy is wasted in the form of heat and sound,5 furthermore the ability to utilize fuels such as
hydrogen5 or hydrazine5,6 allow carbon emissions to be reduced. Although fuel cells have these
advantages over current energy technologies, the challenge with fuel cells has been competing on
a cost, power, and durability basis.4,5,7 These three challenges have guided fuel cell research over
the past decades, and catalysis is at the center of it. With better and cheaper catalysts, costs can
be cut, power density can be improved, and the benefits of fuel cell technologies can be realized.

1!!
!

!

Figure 1.1: Graphical Abstracts of (a) Chapter 2 Carbon Supports for Non-precious Metal
Oxygen Reducing Catalysts, (b) Chapter 3 Solid Alkalinity of MNC Catalyst for Oxygen Reduction
Reaction, (c) Chapter 4 Analysis of Adsorption Effects on a Metal-Nitrogen-Carbon Catalyst using
a Rotating Ring-Disk Study, and (d) Chapter 5 Transport Model for Thick PEMFC Cathodes

1.2. Overview of Work
This dissertation has increased understanding of MNC catalyst function and synthesis with
four chapters:
1. Evaluation of the impact of mesoporosity on catalyst properties: A correlation between
mesoporosity and activity is found, suggesting that mesoporosity increases active site
creation or utilization.

2!!
!

!

2. Exploration of surface alkalinity in MNC catalysts: A correlation between activity and
pH is observed, and the basic site strength and quantity are calculated by two different
methods.
3. Exploration of peroxide generation of MNC catalysts: rotating ring-disk electrode
(RRDE) studies are used to detect peroxide generation and deduce oxygen adsorption
limitations for the MSU catalyst.
4. Numerically model catalyst function in fuel cell context: A thick, hydrophilic fuel cell
membrane-electrode assemblies (MEAs) has been modeled to understand transport in an
MNC catalyst developed by collaborators at the University of New Mexico.
Chapters one and three resulted in peer-reviewed journal publications.8,9 Chapter two was
published as a conference proceeding,10 and Chapter four was presented at the ECS/SMEQ Joint
International Meeting on October, 9th 2014, and the resulting paper is submitted. The final
chapter will consider the impact of the current dissertation, and how future research efforts
should be concentrated.
Of the competing MNC catalyst surveyed in the previous section, these chapters explore two
of them. The first three chapters consider a high-pressure synthesis process developed at MSU in
which the nitrogen precursors break down during the pyrolysis and the high partial pressure of
nitrogen intermediates enhances catalyst site creation. The fourth and final chapter considers the
modeling of a catalyst developed by collaborators at the University of New Mexico (UNM). This
catalyst involves a hard templating of the catalyst particles around silica agglomerates. The
templating ensures high quantities of mesoporosity. The importance of mesoporosity was main
conclusion from chapter two.

3!!
!

!

1.3. Competing Fuel Cell Technologies
A fuel cell is a steady-state galvanic electrochemical reactor in which a fuel is oxidized and
an oxidant reduced in order to convert chemical energy to electrical energy. There are many
different types of fuel cells. They operate at range of temperatures and a variety of fuels, but the
basics are: an anode where the fuel is oxidized, a cathode where the oxidant (generally oxygen)
is reduced, and an electrolyte that allows a charge carrying ion to pass from one electrode to the
other. Table 1.1 shows the various types of fuel cells with the two half-cell reactions, the chargecarrying ions, and the operating conditions.
Fuel Cell Type

Table 1.1: Fuel Cell Types
Cathode
Charge-Carrying Ion
Reaction
(Electrolyte)
! "
+

Proton Exchange
Membrane Fuel
Cell (PEMFC)

O2 +
+
4H +4e
"H2O

H (in a polymer
membrane)

Alkaline Fuel
Cell (AFC)

O2+2H2O+4
e "4OH

OH

Solid Oxide Fuel
Cell (SOFC)

O2+2e "
2O

O

Phosphoric Acid
Fuel Cell (PAFC)

O2 +
+
4H +4e
"H2O
O2+2CO2+4
2e " 2CO3

H (in phosphoric
acid)

Molten
Carbonate Fuel
Cell (MCFC)

-

-

2-

+

CO3

2-

4,5

Anode
Reaction

Operating
Conditions

+

2H2"4H +
4e

-

2H2+4OH
"4H2O+4
e
2O
+Fuel"2e
+Fuel
oxide
+
2H2"4H +
4e
2H2+
22CO3
"4H2O+2
CO2+4e

50-100 °C with air
and hydrogen
4
gases
60-200 °C with air
and hydrogen
5
gases
600-1000 °C with
air and a variety
4
of fuels
200 °C with air
and hydrogen
4
gases
650 °C with air
and hydrogen
4
gases

Of these fuel cells, solid oxide fuel cells (SOFCs), phosphoric acid fuel cells (PAFCs), and
molten carbonate fuel cells (MCFCs) are considered high temperature fuel cells and are used
primarily in stationary applications, often when combined heat and power (CHPs) is desired.4,5
Proton exchange membrane fuel cells (PEMFCs) and alkaline fuel cells (AFCs) are considered

4!!
!

!

low-temperature fuel cells and are the most researched fuel cells for transportations
applications.4,5
AFCs have been widely used in the space industry and have the advantage of relatively facile
kinetics in comparison to other fuel cell systems, but degradation issues have been the most
difficult challenge.4,5,11,12 Specifically, in aqueous AFCs, any carbon dioxide in the system reacts
with hydroxide to form carbonate and bicarbonate, which decreases the electrolyte pH.4,11,12
Because OH- is the charge-carrying ion, any decrease in pH results in a decrease in conductivity.
Furthermore, these carbonate and bicarbonate ions can degrade electrode material. 4 For this
reasons much research has been devoted to developing anion exchange membranes (AEM).13
Although these membranes improve carbon dioxide tolerance, they generally have stability
issues especially at lower hydration levels.4,13 Although much research has been spent on
improving stability, this challenge has caused the transportation industry to invest most of their
resources towards PEMFCs.7,14
PEMFCs have also been used in the space industry and are the most widely considered fuel
cell for automotive applications.4,5,14 This is due to the high durability and high conductivity of
the ionomer membrane.4,5 The challenge with the PEMFC has been that both half-cell reactions
require precious metal catalysts.4,5 The present dissertation concerns this challenge.
As mentioned in the overview, the primary challenges in competing with internal combustion
engines on a cost and performance basis. The primary impediment to competing on a cost basis
is the expense of the catalysts. This loading of costly catalysts must also be balanced with the
competing concerns of power and durability. Of the two half-cell reactions, the oxygen reduction
reaction (ORR) is the most sluggish resulting higher platinum loadings. The anode platinum
loadings are generally on the order of 0.05 mg/cm2, while the cathode loadings range from 0.2-

5!!
!

!

0.4 mg/cm2.4 Due to the higher platinum loading, the cathode catalyst has been the focus of the
most research.4 Table 1.2 shows various catalysts that have been considered for ORR in
PEMFCs.
Table 1.2: Oxygen Reduction Catalysts
Catalyst

Performance at
-2
0.6 V / A cm

General
Form

Metals

Non-metals

Platinum and Platinum Alloys

1.5

15

PtxMey

Me=Ni, Mn, Fe,
16
Co, Cu

C-support

Metal-Nitrogen-Carbon
HT catalyst

0.4

17

MeNxCy

Me=Fe, Co,
18
Ni, Cu, Mn

N,C

Chalcogenides
Transition Metal Nitrides and
Carbides

0.18

19

MemX8

Me=Mo, Ru

X=Se

0.05

20

MeXx

Me=Ti,Ta,W,Mo

X=N,C

For precious metal catalysts the goal is to reduce Pt usage without effecting durability or
activity. The current state-of-the-art catalyst for ORR in PEMFCs is de-alloyed PtNi3.15,21,22 The
catalyst is based on a precursor of PtNi3 alloy nanoparticles. Upon heat treatment, de-alloying
causes the surface of the nanoparticles to be platinum rich, while the centers are primarily
composed of the cheaper Ni metal.15,21,22 This catalyst has been found to function better the pure
platinum not only on a Pt-mass basis, but also the interior Ni has been found to enhance catalytic
properties of the Pt-rich surface.15,21,22 Although these catalyst have extremely high activity on a
platinum basis (0.35 A/mgPt), this level of Pt use is still too high (2015 goal: 0.44 A/mgPt).23,24
As shown in Table 1.2, there are a number of alternatives to platinum group metal (PGM)
catalysts for ORR. The most competitive is metal nitrogen carbon (MNC) catalysts. Jasinski
discovered the catalytic activity of metal-N4 macrocycles.25 Later developments by Yeager et al.
showed improved catalytic activity and durability from a pyrolyzed mixture of cobalt acetate,
carbon, and polyacrylonitrile.26 Ultimately this classification consists of a wide variety of
catalysts, but the synthesis procedure generally involves a mixing step involving metal, nitrogen,
and carbon precursors followed by a pyrolysis step at 700-1000 °C.8,17,27-31 The initial pyrolysis
6!!
!

!

step is followed by a leaching step to remove excess metal that can be harmful to Nafion and
often a second or third pyrolysis.17,27,31 These MNC catalysts can be broken down into three main
categories: ammonia activated catalysts,32,33 autogenic pressure generating catalysts,34,35 and
high-molecular-weight nitrogen precursor catalysts.31,36 The ammonia activated catalysts are
synthesized by purging the reactor with ammonia gas during the first pyrolysis. Normally these
catalysts have a second pyrolysis under an inert atmosphere to increase durability. The autogenic
pressure generating catalysts use nitrogen precursors that decompose during pyrolysis. The
decomposition of the nitrogen precursors creates a high pressure of gaseous intermediates that
react with the remaining solid phase to create active sites.34,35,37 These catalysts were developed
at Michigan State University and are the subject of Chapters 2-4. High molecular-weight
nitrogen precursor catalysts are synthesized with nitrogen or metal-nitrogen precursors such as
nicarbazim,32 polyaniline,31 or metal macrocycles.38,39 These catalysts can be synthesized using
an atmospheric-pressure, purged reactor without losing enough nitrogen precursor to hurt
activity. One of these catalysts was developed at the University of New Mexico32 and will be the
subject of Chapter 5.

7!!
!

!

Figure 1.2: Synthesis of MNC catalysts

The current state of the art MNC catalyst is a high-molecular-weight MNC developed by
Zelenay et al at Los Alamos National Laboratories.31

For this catalyst, the carbon support, iron-

salt, and short chain aniline oligomer are mixed. After the mixing, an oxidant is added to
polymerize the aniline. The polymerization step is followed by a heat treatment in nitrogen at
900 °C, a leaching step in sulfuric acid, and a second heat treatment in nitrogen atmosphere.31
1.4. Methods and Fundamentals
1.4.1

Physical Characterization

The primary surface morphology techniques used in this dissertation consider porosity. We
will consider porosimetry techniques based on nitrogen physisorption and mercury intrusion.
The nitrogen physisorption analysis methods used in this dissertation are: non-local density
functional theory (NLDFT); the Brunauer, Emmett, and Teller (BET) method (a method for
8!!
!

!

establishing specific surface area of a substance);40 and the Barrett, Joyner, and Halenda (BJH)
method (a method for establishing pore size distributions for pores larger than 2 nm).41

!
Figure 1.3: Typical Nitrogen Isotherm showing both adsorption and desorption curves on Ketjen
EC 600J carbon black at 77 K.

All three nitrogen physisorption techniques are calculations based on adsorption isotherms.
Figure 1.3 shows a typical isotherm of Ketjen EC 600J, a material commonly used in the
synthesis of MNC catalysts. As the nitrogen pressure is increased at a constant temperature
(generally 77 K), nitrogen molecules adsorb to the sample surface as the saturation pressure is
approached. The BET surface area is based on a calculation of the volume of a single layer of
nitrogen atoms on the sample surface, a monolayer. Essentially, it is assumed that molecules
adsorb to the surface with one heat of adsorption, and that molecules adsorb to other adsorbed
molecules with a different heat of adsorption. Calculation of equilibrium surface coverage based
on Arrhenius adsorption laws as well as significant algebraic manipulation results in the BET
equation:41

9!!
!

!

P/P0
1
(c 1)P/P0
=
+
Q(1 P/P0 )
cvm
cvm

[1]

where P is measured pressure, P0 is saturation pressure, Q is measure adsorbed volume, vm is
monolayer volume, and c is a constant representing the ratio of adsorption energies for the first
monolayer (adsorbent to surface) to the remaining monolayers (adsorbent to adsorbent). The lefthand side of Eq. 1 can be regressed with respected to P/P0 in order to deduce vm from the slope
(SBET) and intercept (JBET):

vm =

1
SBET + JBET

[2]

Then, given the size of the nitrogen atoms, a surface area can be calculated from the monolayer
volume.41
BJH is a method of determining the pore size distribution over a relatively wide range
spanning primarily the mesoporous region (~2-50nm).42 In this method, the pore structure is
modeled as a system of cylindrical pores. The pores associated with a specific diameter can be
considered a single cylindrical pore with a given length. The method also assumes an initial
condition where all pores are filled (i.e. at saturation pressure). As the pressure decreases, the
critical pressures of the liquid phase nitrogen in the center of the pores are reached. Essentially,
smaller, more confined pores allow the liquid nitrogen to be more stable decreasing the critical
pressure. In this way, each decreasing pressure value can be associated with a certain pore radius.
When the decreasing pressure reaches the critical value for a given pore size, the vaporization of
the liquid nitrogen in the pore causes a decrease in adsorbed nitrogen shown in the desorption
curve in Figure 1.3. BJH theory uses the Kelvin equation to model the relationship between
observed vapor pressure, P, and pore of radius, r: 40

10!
!
!

!

RT ln(P/P0 ) =

V
r

[3]

where R is the ideal gas constant, T is the absolute temperature, P0 is the normal vapor pressure,
V is molar volume, and

is the surface tension. In this way desorption between two different

pressures can be associated with vaporization of liquid in pores or a certain radius allowing the
calculation of a pore size distribution.40
Surface area and pore size distribution information are also obtained using a non-local
density functional theory (NLDFT) model for non-graphitic carbons developed by Ustinov.43,44
In this approach, density functions for various pore sizes and pressures were obtained by
minimizing the grand thermodynamic potential, Ω, as a function of local density. ρ:
Z
⌦ = ⇢[f (x, ⇢, ⇢m ) + uext (x) µ]dx

[4]

where uext is the external potential exerted by the solid, µ is the chemical potential, and f is
the local molar Helmholtz free energy. The Helmholtz free energy is a function of position, x;
local density, ρ; and smoothed density, ρm. Ustinov calculated molecular parameters from
nitrogen adsorption on a non-graphitized carbon black (BP 280) and incorporated these
parameters in the energy equation to find density as a function of pressure. The minimization of
Eq. for various pore sizes and pressures, allows a mean density to be calculated. This kernel of
density as it varies with pressure and pore size can be used to deconvolute an adsorption
isotherm into adsorption associated with specific pore sizes. The total adsorption, Q as a function
of pressure, P, is:

Q(P ) =

X
i

11!
!
!

⇢m,i |P vi

[5]

!

where ρm,i|P is the mean density in the ith pore at pressure, P, and vi is volume of the ith pore.
The set of v is the discretized pore size distribution.
So far the porosity techniques have only considered pores smaller than about 300nm
diameter. Larger pores are generally measured with mercury intrusion porosimetry (MIP). In this
method, mercury is forced into pores by increasing pressure. As the pressure increases, mercury
is forced into increasingly smaller pores. This phenomenon is governed by the surface energy of
mercury via by Washburn’s equation:45

rc =

2 cos(✓)
Pc

[6]

where Pc is the mercury pressure, γ is the surface tension, theta is the contact angle, and rc is the
pore radius. Using the consistent surface properties of mercury a pore size distribution is
calculated.
1.4.2

Electrochemical Characterization

A number of electrochemical methods will be useful for understanding this dissertation:
linear sweep voltammetry (LSV), cyclic voltammetry (CV), rotating disc electrode (RDE), and
rotating ring-disc electrode (RRDE). Both LSV and CV are electrochemical methods in which
the potential is varied and the current is allowed to respond using a three electrode system.46 A
three-electrode system consists of a working, counter, and reference electrode. The reaction of
interest occurs at the working electrode, while a second reaction is required for charge neutrality
at the counter electrode. To eliminate the impact of ohmic and kinetic potential drops across the
counter electrode, a third electrode is required. The third electrode is the reference: it contains a
known electrochemical reaction. Measuring the potential of the working electrode with respect to
the equilibrium potential of the reference electrode reaction eliminates the impact of the counter

12!
!
!

!

electrode. The main difference between LSV and CV is that CV consists of multiple potential
sweeps between two potentials. A typical voltammetry curve consists of three main sections. If
one starts at open circuit (on the right side of Fig. 1.2), it begins with a section in which there is
no faradaic current because the electric field at the electrode surface is not large enough to drive
the reaction. Eventually, as the potential sweeps farther from open circuit (middle of Fig. 1.2),
the electric field at the electrode surface becomes large enough to drive the reaction. Initially this
over-potential is small and results the reaction being limited by the kinetics of the
electrochemical reaction. Eventually the reaction rate increases to the extent that the electrode
surface becomes depleted of reactant molecules (left side of Fig. 1.2). At this point the reaction
becomes limited by the ability of the reactant to diffuse to the electrode surface. As the reactant
is depleted from the surface, the surface concentration approaches zero. When the surface
concentration is essentially zero, the reaction rate is governed by the rate at which the reactant
can diffuse to the surface, and the current becomes independent of potential. These three regions
will now be covered in detail.

13!
!
!

!

!
Figure 1.4: Typical cyclic voltammogram for a reduction reaction.

Although the primary parameter of interest for electrocatalysts is current, the open circuit
potential can also give information about the catalyst. The open circuit potential is the result of
equilibrium between the forward and reverse reactions (for this work the forward reaction is a
reduction reaction and the reverse reaction is oxidation) resulting in zero net current. This
equilibrium is considered using Gibbs free energy and results in the Nernst Equation. If we
consider currents to be reversible, we can express a change in Gibbs free energy, G, as a charge
passed over a reversible potential, E:46
G=

nF E

[7]

where n is the number of electrons per molecule, and F is Faraday’s constant representing charge
per mole electron. A reference Gibbs free energy is defined This is generally defined as a
standard free energy and reversible potential when all substances are 1 M. Activity is measure of

14!
!
!

!

chemical availability and is generally considered to be to molarity. When substances are no
longer present at unit activity, the chemical potential can be considered:

G=

G0 + RT ln(a)

[8]

where the superscript 0 refers to standard values, R is the ideal gas constant, T is the temperature,
and a is activity. Activity is the availability of a constituent relative to 1 M concentration or 1
atmosphere partial pressure. A reference Gibbs free energy is defined when all constituents have
unit activity. In terms of the potential this becomes the Nernst Equation:46

E = E0 +

RT
aox
ln(
)
nF
ared

[9]

where the potential is a function of the ideal gas constant, R; the temperature, T; and the activity
of the oxidant and reductant, aox and ared respectively.
When the polarization curve moves to non-equilibrium regions, a new kinetic model is used.
In these cases, one of the most general kinetic models is Butler-Volmer. This model assumes that
forward and reverse reaction rates follow an Arrhenius form: 46

k = A exp (

G/RT )

[10]

where the reaction rate, k, is a function of some pre-exponential, A, and some energy barrier, ΔG.
Considering Eq. 7 and 9, these reaction rates can be considered in terms of potential: 46

k = k0 exp ((E 0

E)↵F/RT )

where k0 is some equilibrium rate constant, and α is an exchange coefficient that represents
some fraction of the activation barrier. When we consider both an oxidation and reduction
reaction rate constant, the rate, expressed as a current is the Butler-Volmer Equation: 46

15!
!
!

[11]

!

i = nF k0 [Cox exp ( ↵(E

E 0 )F/RT )

Cred exp ((1

↵)(E

E 0 )F/RT )]

[12]

where the current density, i, is a function of a rate constant, k0 and concentrations at the surface
for oxidant, Cox, and reductant, Cred. Note that the nF converts the rate to a current. Often it is
assumed that the reaction is symmetrical, meaning that the potential deviation equally affects the
anodic and cathodic branches resulting in an exchange coefficient is 0.5. An approximation of
this kinetic model is Tafel kinetics in which it is assumed one branch is much slower than the
other branch, that is:

Cox exp ( ↵f (E

E 0 ))

Cred exp ((1

↵)f (E

E 0 ))

[13]

or vise versa. In this case, the Butler-Volmer expression simplifies to: 46

i = nF k0 C exp ( ↵f (E

E 0 ))

[14]

This equation is the Tafel equation, and results in an exponential relationship between potential
and current. This Tafel relationship allows for quick and easy identification of kinetically limited
regions of polarization curves.
A second aspect of these electrochemical characterization techniques is how the basic
technique (LSV, CV, etc.) is incorporated applied to specific chemistry, electrode design, and
transport system. For this dissertation, the chemistry will generally be oxygen reduction, and
thorough discussion is included in Chapter 4. The oxygen reduction reaction (ORR) can follow
two different pathways (Fig. 1.5).47-52 The first pathway is direct four-electron reduction to water
without a desorbing intermediate (k1 in Fig. 1.5). The second pathway is incomplete reduction of
oxygen to hydrogen peroxide (k2). The hydrogen peroxide can desorb from the surface (KP,2),47
disproportionate to water and oxygen,49 or reduce to water (k3).49

16!
!
!

!

Figure 1.5: Reaction schematic showing two pathways: pathway one (Ko,1 and k1) is the complete
reduction of oxygen to water with no desorbing intermediate and pathway two (Ko,2 and k2) is the twoelectron reduction to a desorbing peroxide intermediate that disperses into the catalyst layer (Kp,2), but
also has the potential to be further reduced to water (k3).

To return to the explanation of electrochemical characterization techniques, one commonly
used system is rotating disk electrode (RDE). As shown in Figure 1.6, the system consists of a
rotating, glassy-carbon, working electrode with a layer of catalyst on it, a reference electrode,
and platinum counter electrode in an acidic solution. Oxygen is bubbled into the solution to
ensure that the bulk oxygen concentration in the electrolyte remains at saturation. The protons
are supplied by the acidic solution, but charge neutrality is conserved due to the reaction at the
counter electrode. Essentially, the drag caused by the rotation of the disk induces a radial flow of
electrolyte at the surface. The centrifugal acceleration causes a flux radially away from the disk
center and axially towards the disk as shown in Fig. 1.6. In this way the rotation of the working
electrode allows LSV or CV with controllable diffusion limitations.46,53,54 Rotating ring-disk
electrodes (RRDEs) are also used extensively in this work. In the RRDE system, a second
working electrode is used. This fourth electrode is a concentric ring around the disk, and any by17!
!
!

!

products resulting from reactions at the disk pass the ring electrode due to flow field in the
system.46,47

Figure 1.6: Rotating Ring-Disc Electrode Experimental Set-up

Due to the complexity involved in Chapter 4, this introduction will cover the transport
environment surrounding RDEs and RRDEs in more detail. The diffusion limitations are well
known via the Levich equation.53 Although a complete derivation of the Levich equation is
outside of the scope of this introduction, an explanation and discussion will suffice. The full
derivation is well documented.53,54 Initially, the flow field around the disk is calculated from a
solution to the Navier-Stokes equations in the electrolyte surrounding the disk. It is assumed that
there is a no-slip boundary condition at the electrode surface and that the electrolyte is stagnant
and large distances from the disk. The hydrodynamic boundary layer has a thickness, δ0:53,54
18!
!
!

!

0

p
= 3.6 ⌫/!

[15]

Using the flow field and solution to the convective diffusion equation can be found. The
boundary conditions consist of some concentration at the surface and some bulk concentration at
distances far from the disk. This solution allows the concentration to be written as a function of
velocity and position and results in a diffusion boundary layer, δD:53,54
D

= 1.61D1/3 ⌫ 1/6 !

1/2

[16]

where D is a diffusivity, ν is viscosity, and ω is rotation speed. A comparison of the
hydrodynamic and diffusion boundary layers shows that the diffusion layer thickness is about
6% of the hydrodynamic layer thickness for oxygen in water (viscosity of 0.01 cm2/s and oxygen
diffusivity of 2 ×10-5 cm2/s): 55,56
D

= 0.45

0

D1/3
⌫ 1/3

[17 ]

The relative thinness of the diffusion boundary layer ensures that diffusion dominates mass
transport near the electrode surface. Assuming Fickian diffusion near the surface, a relationship
between current, i, and concentration at the surface, Cs can be found:47,48,51
i = nF D

(Cs

Cb )

[18]

D

where Cb is the bulk concentration. This relationship is the foundation for Chapter 4, and allows
the R(R)DE system to be useful in isolating a catalyst’s performance from transport
considerations. By setting the surface concentration to zero and combining Eqs. 16 and 18, the
Levich equation is obtained:47,48,51
imt = 0.62nF D2/3 ⌫
19!
!
!

1/6

! 1/2 Cb

[19]

!

The Levich equation calculates the mass transfer limiting current, imt, in a RDE system.
The second important transport system in this dissertation is a membrane electrode assembly
(MEA). This transport system is important to the oxygen reduction reaction because it represents
the intended application of these catalysts. The MEA system, shown in Fig. 1.1d, allows the
transport of protons and oxygen to active sites in a conductive solid network. A further transport
consideration is the egress of water from the MEA. The transport of protons as well as the
electric current network are generally considered using Ohm’s law in which the current, i, is
proportional to the gradient of the electric field:57

i = rV

[20]

where V is potential and κ is conductivity. The proton flux represents a current through the
electrolyte phase of the MEA. Besides ohmic considerations there are also mass transport
considerations for gas and liquid species. In the gas phase, multi-component diffusion must be
considered due to the presence of oxygen, nitrogen, and water vapor. The flux of these
components is dependent on their concentrations and is often considered with Stefan-Maxwell
diffusion which accounts for momentum conservation of randomly colliding gas molecules:57

rxi =

X Ni xj

Nj x i

cDij

i6=j

[21]

where the vapor fraction of constituent i, xi, is related to the fluxes, N, and gas phase
concentration, c, as well as binary diffusivity between constituent i and j, Dij. The liquid phase is
generally considered pure water, and the Hagen-Poiseuille equation relates the liquid flux, NL to
the pressure gradient:57
NL =

k
rPL
µ
20!
!

!

[22]

!

where k is permeability, µ is viscosity, and PL is pressure in the liquid phase. Because
understanding the transport requires the deconvolution of the impact of four different phases,
MEA results are not necessarily very useful in understanding catalyst kinetics. This complexity
is why RDE results are so useful as shown in Chapter 4. On the other hand, the importance of the
MEA results to catalyst application requires full understanding. This MEA transport
understanding is the goal of Chapter 5.
!

21!
!
!

!
!
!
!
!
!
!
!
!
!
!
!
!
REFERENCES

22!
!
!

!

REFERENCES

1.

International Energy Outlook 2014, (ed. Doman, L.), U.S. Energy Information
Administration, DOE/EIA-0484(2014) (2014). <http://www.eia.gov/forecasts/ieo/>

2.

World Population Prospects: The 2012 Revision, Population Division of the Department of
Economic and Social Affairs of the United Nations Secretaria, (2012). <
http://esa.un.org/unpd/wpp/index.htm >

3.

Mitigation of Climate Change: Technical Summary, Intergovenmental Panel on Climate
Change, (2014).

4.

K.-D. Kreuer, Fuel Cells: Selected Entries from the Encyclopedia of Sustainability Science
and Technology,(Springer, 2012).

5.

J. Larminie and A. Dicks, Fuel cell systems explained,(J. Wiley, New York, 2003).

6.

A. Serov and C. Kwak, "Direct hydrazine fuel cells: A review," Applied Catalysis B:
Environmental, 98(1), 1-9 (2010).

7.

M. K. Debe, "Electrocatalyst approaches and challenges for automotive fuel cells," Nature,
486(7401), 43-51 (2012).

8.

N. Leonard, V. Nallathambi and S. C. Barton, "Carbon Supports for Non-Precious Metal
Oxygen Reducing Catalysts," Journal of the Electrochemical Society, 160(8), F788-F792
(2013). doi:10.1149/2.026308jes

9.

N. D. Leonard and S. C. Barton, "Analysis of Adsorption Effects on a Metal-NitrogenCarbon Catalyst Using a Rotating Ring-Disk Study," Journal of The Electrochemical
Society, 161(13), H3100-H3105 (2015).

10. N. D. Leonard, S. Ganesan, V. Nallathambi and S. C. Barton, "Alkalinity of Non-Precious
Metal Catalysts for Oxygen Reduction," ECS Transactions, 50(2), 1869-1873 (2013).
11. J. R. Varcoe and R. C. Slade, "Prospects for Alkaline Anion
Temperature Fuel Cells," Fuel cells, 5(2), 187-200 (2005).

Exchange Membranes in Low

12. G. F. McLean, T. Niet, S. Prince-Richard and N. Djilali, "An assessment of alkaline fuel
cell technology," International Journal of Hydrogen Energy, 27(5), 507-526 (2002).
doi:10.1016/S0360-3199(01)00181-1
13. J. R. Varcoe, P. Atanassov, D. R. Dekel, A. M. Herring, M. A. Hickner, P. A. Kohl, A. R.
Kucernak, W. E. Mustain, K. Nijmeijer, K. Scott, T. Xu and L. Zhuang, "Anion-exchange
membranes in electrochemical energy systems," Energy & Environmental Science, 7(10),
3135-3191 (2014). doi:10.1039/C4EE01303D
23!
!
!

!

14. "FUEL CELL R&D ACTIVITIES," http://energy.gov/eere/fuelcells/fuelDcellDrdDactivities,
accessed 10/28.
15. B. Han, C. E. Carlton, A. Kongkanand, R. S. Kukreja, B. R. Theobald, L. Gan, R. O'Malley,
P. Strasser, F. T. Wagner and Y. Shao-Horn, "Record activity and stability of dealloyed
bimetallic catalysts for proton exchange membrane fuel cells," Energy & Environmental
Science (2014).
16. C. Wang, N. M. Markovic and V. R. Stamenkovic, "Advanced Platinum Alloy
Electrocatalysts for the Oxygen Reduction Reaction," ACS Catalysis, 2(5), 891-898 (2012).
doi:10.1021/cs3000792
17. F. Jaouen, J. Herranz, M. Lefevre, J.-P. Dodelet, U. I. Kramm, I. Herrmann, P. Bogdanoff, J.
Maruyama, T. Nagaoka, A. Garsuch, J. R. Dahn, T. Olson, S. Pylypenko, P. Atanassov and
E. A. Ustinov, "Cross-Laboratory Experimental Study of Non-Noble-Metal Electrocatalysts
for the Oxygen Reduction Reaction," ACS Applied Materials & Interfaces, 1(8), 1623-1639
(2009). doi:10.1021/am900219g
18. E. Yeager, "Electrocatalysts for O-2 Reduction," Electrochimica Acta, 29(11), 1527-1537
(1984). doi:10.1016/0013-4686(84)85006-9
19. Y. Feng, A. Gago, L. Timperman and N. Alonso-Vante, "Chalcogenide metal centers for
oxygen reduction reaction: Activity and tolerance," Electrochimica Acta, 56(3), 1009-1022
(2011). doi:10.1016/j.electacta.2010.09.085
20. H. Peng, Z. Mo, S. Liao, H. Liang, L. Yang, F. Luo, H. Song, Y. Zhong and B. Zhang,
"High Performance Fe- and N- Doped Carbon Catalyst with Graphene Structure for Oxygen
Reduction," Sci. Rep., 3 (2013). doi:10.1038/srep01765
21. F. Hasché, M. Oezaslan and P. Strasser, "Activity, structure and degradation of dealloyed
PtNi3 nanoparticle electrocatalyst for the oxygen reduction reaction in PEMFC," Journal of
The Electrochemical Society, 159(1), B24-B33 (2011).
22. P. Strasser, S. Koh, T. Anniyev, J. Greeley, K. More, C. Yu, Z. Liu, S. Kaya, D. Nordlund,
H. Ogasawara, M. F. Toney and A. Nilsson, "Lattice-strain control of the activity in
dealloyed core-shell fuel cell catalysts," Nat Chem, 2(6), 454-60 (2010).
doi:10.1038/nchem.623
23. Nanosegregated Cathode Alloy Catalysts with Ultra-Low Platinum Loading, DOE
Hydrogen and Fuel Cells Program, (2013).
24. DOE Hydrogen Program- 2008 Annual Progress Report, United States Department of
Energy, (2008). <http://www.hydrogen.energy.gov/annual_progress08.html>
25. R. Jasinski, "A New Fuel Cell Cathode Catalyst," Nature, 201(4925), 1212-1213 (1964).

24!
!
!

!

26. S. Gupta, D. Tryk, I. Bae, W. Aldred and E. Yeager, "Heat-Treated Polyacrylonitrile-Based
Catalysts for Oxygen Electroreduction," Journal of Applied Electrochemistry, 19(1), 19-27
(1989).
27. F. Jaouen, E. Proietti, M. Lefèvre, R. Chenitz, J.-P. Dodelet, G. Wu, H. T. Chung, C. M.
Johnston and P. Zelenay, "Recent advances in non-precious metal catalysis for oxygenreduction reaction in polymer electrolyte fuel cells," Energy & Environmental Science, 4(1),
114 (2011). doi:10.1039/c0ee00011f
28. H. T. Chung, C. M. Johnston, K. Artyushkova, M. Ferrandon, D. J. Myers and P. Zelenay,
"Cyanamide-derived non-precious metal catalyst for oxygen reduction," Electrochemistry
Communications, 12(12), 1792-1795 (2010). doi:10.1016/j.elecom.2010.10.027
29. A. Serov, M. H. Robson, M. Smolnik and P. Atanassov, "Templated bi-metallic non-PGM
catalysts for oxygen reduction," Electrochimica Acta, 80(0), 213-218 (2012).
doi:10.1016/J.Electacta.2012.07.008
30. V. Nallathambi, N. Leonard, W. Patterson, K. Artyushkova, P. Atanassov and S. C. Barton,
"Ammonia-Generating Precursors in MNC Electrocatalysts for Oxygen Reduction,"
Meeting Abstracts(16), 1150-1150 (2011).
31. R. Bashyam and P. Zelenay, "A class of non-precious metal composite catalysts for fuel
cells," Nature, 443(7107), 63-66 (2006). doi:10.1038/nature05118
32. E. Proietti, F. Jaouen, M. Lefevre, N. Larouche, J. Tian, J. Herranz and J. P. Dodelet, "Ironbased cathode catalyst with enhanced power density in polymer electrolyte membrane fuel
cells," Nature communications, 2, 416 (2011). doi:10.1038/ncomms1427
33. M. Lefevre, E. Proietti, F. Jaouen and J. P. Dodelet, "Iron-Based Catalysts with Improved
Oxygen Reduction Activity in Polymer Electrolyte Fuel Cells," Science, 324(5923), 71-74
(2009). doi:10.1126/science.1170051
34. R. Kothandaraman, V. Nallathambi, K. Artyushkova and S. C. Barton, "Non-precious
oxygen reduction catalysts prepared by high-pressure pyrolysis for low-temperature fuel
cells,"
Applied
Catalysis
B
Environmental,
92(1-2),
209-216
(2009).
doi:10.1016/J.Apcatb.2009.07.005
35. V. Nallathambi, J. W. Lee, S. P. Kumaraguru, G. Wu and B. N. Popov, "Development of
high performance carbon composite catalyst for oxygen reduction reaction in PEM Proton
Exchange Membrane fuel cells," Journal of Power Sources, 183(1), 34-42 (2008).
doi:10.1016/j.jpowsour.2008.05.020
36. A. Serov, K. Artyushkova and P. Atanassov, "Fe-N-C Oxygen Reduction Fuel Cell Catalyst
Derived from Carbendazim: Synthesis, Structure, and Reactivity," Advanced Energy
Materials, 4(10), n/a-n/a (2014). doi:10.1002/aenm.201301735

25!
!
!

!

37. V. Nallathambi, N. Leonard, R. Kothandaraman and S. Calabrese Barton, "Nitrogen
Precursor Effects in Iron-Nitrogen-Carbon Oxygen Reduction Catalysts," Electrochemical
and Solid State Letters, 14(6), B55-B58 (2011). doi:10.1149/1.3566065
38. S. Maldonado and K. J. Stevenson, "Direct preparation of carbon nanofiber electrodes via
pyrolysis of iron(II) phthalocyanine: Electrocatalytic aspects for oxygen reduction," Journal
of Physical Chemistry B, 108(31), 11375-11383 (2004). doi:10.1021/Jp0496553
39. P. Bogdanoff, I. Herrmann, M. Hilgendorff, I. Dorbandt, S. Fiechter and H. Tributsch,
"Probing structural effects of pyrolysed CoTMPP-based electrocatalysts for oxygen
reduction via new preparation strategies," Journal of New Materials for Electrochemical
Systems, 7(2), 85-92 (2004).
40. S. Brunauer, P. H. Emmett and E. Teller, "Adsorption of Gases in Multimolecular Layers,"
Journal
of
the
American
Chemical
Society,
60(2),
309-319
(1938).
doi:10.1021/ja01269a023
41. E. P. Barrett, L. G. Joyner and P. P. Halenda, "The Determination of Pore Volume and Area
Distributions in Porous Substances. I. Computations from Nitrogen Isotherms," Journal of
the American Chemical Society, 73(1), 373-380 (1951). doi:10.1021/ja01145a126
42. F. Jaouen, S. Marcotte, J.-P. Dodelet and G. Lindbergh, "Oxygen Reduction Catalysts for
Polymer Electrolyte Fuel Cells from the Pyrolysis of Iron Acetate Adsorbed on Various
Carbon Supports," The Journal of Physical Chemistry B, 107(6), 1376-1386 (2003).
doi:10.1021/jp021634q
43. E. A. Ustinov, D. D. Do and V. B. Fenelonov, "Pore size distribution analysis of activated
carbons: Application of density functional theory using nongraphitized carbon black as a
reference system," Carbon, 44(4), 653-663 (2006). doi:10.1016/j.carbon.2005.09.023
44. E. A. Ustinov, "Nitrogen Adsorption on Silica Surfaces of Nonporous and Mesoporous
Materials," Langmuir, 24(13), 6668-6675 (2008). doi:10.1021/la704011z
45. A. W. Adamson and A. P. Gast, Physical chemistry of surfaces,(Wiley, New York, 1997).
46. A. J. Bard and L. R. Faulkner, Electrochemical methods : fundamentals and
applications,(John Wiley, New York, 2001).
47. N. A. Anastasijevic, V. Vesovic and R. R. Adzic, "Determination of the kinetic parameters
of the oxygen reduction reaction using the rotating ring-disk electrode: Part I. Theory,"
Journal of Electroanalytical Chemistry and Interfacial Electrochemistry, 229(12), 305-316
(1987). doi:10.1016/0022-0728(87)85148-3
48. N. Anastasijević, V. Vesović and R. Adžić, "Determination of the kinetic parameters of the
oxygen reduction reaction using the rotating ring-disk electrode: Part II. Applications,"
Journal of electroanalytical chemistry and interfacial electrochemistry, 229(1), 317-325
(1987).
26!
!
!

!

49. F. Jaouen, "O2 Reduction Mechanism on Non-Noble Metal Catalysts for PEM Fuel Cells.
Part II: A Porous-Electrode Model To Predict the Quantity of H2O2 Detected by Rotating
Ring-Disk Electrode," The Journal of Physical Chemistry C, 113(34), 15433-15443 (2009).
50. H. S. Wroblowa, P. Yen Chi and G. Razumney, "Electroreduction of oxygen: A new
mechanistic criterion," Journal of Electroanalytical Chemistry and Interfacial
Electrochemistry, 69(2), 195-201 (1976). doi:10.1016/S0022-0728(76)80250-1
51. K. L. Hsueh, D. T. Chin and S. Srinivasan, "Electrode kinetics of oxygen reduction: A
theoretical and experimental analysis of the rotating ring-disc electrode method," Journal of
Electroanalytical Chemistry and Interfacial Electrochemistry, 153(1), 79-95 (1983).
doi:10.1016/S0022-0728(83)80007-2
52. A. Damjanovic, M. A. Genshaw and J. O. M. Bockris, "The Role of Hydrogen Peroxide in
the Reduction of Oxygen at Platinum Electrodes," The Journal of Physical Chemistry,
70(11), 3761-3762 (1966). doi:10.1021/j100883a515
53. V. G. Levich, Physicochemical Hydrodynamics,(Prentice-Hall, Englewood Cliffs, NJ, 1962).
54. Y. V. Pleskov and V. Y. Filinovski, The Rotating Disc Electrode,(Consultants Bureau, New
York, 1976).
55. K. E. Gubbins and R. D. Walker, "The solubility and diffusivity of oxygen in electrolytic
solutions," Journal of The Electrochemical Society, 112(5), 469-471 (1965).
56. W. M. Haynes (ed.) CRC Handbook of Chemistry and Physics (CRC Press, 2010).
57. A. Z. Weber and J. Newman, "Modeling Transport in Polymer-Electrolyte Fuel Cells,"
Chemical Reviews, 104(10), 4679-4726 (2004).

!
!

27!
!
!

Chapter 2 Carbon Supports for Non-precious Metal Oxygen Reducing Catalysts
2.1 Abstract
Porous carbon materials with varying structural and compositional properties were studied for
their impact on the nitrogen content and activity of metal-nitrogen-carbon (MNC) oxygen
reduction catalysts prepared using high-pressure pyrolysis. The carbon materials and resulting
catalysts were characterized morphologically using nitrogen physisorption, coupled with nonlocal density functional theory (NLDFT) analysis to calculate pore size distributions.
Graphiticity was assessed via X-ray Diffraction (XRD), bulk nitrogen content was observed
using CHN combustion analysis and iron content by Inductively Coupled Plasma (ICP). The
catalysts were characterized electrochemically using rotating ring-disk measurements. The
results indicate that substrates adsorbing the most nitrogen and iron show the highest activity.
Furthermore, a relationship found between mesoporosity and nitrogen adsorption indicate the
importance of transport of precursors to potential active sites.
2.2 Introduction
The cost of precious metals has driven the search for lower-cost alternatives to catalyze the
oxygen reduction reaction (ORR) and accelerate commercialization of low-temperature fuel
cells. An important class of non-precious metal catalysts for oxygen reduction is pyrolyzed
metal/nitrogen/carbon (MNC) compounds.1-3 These compounds involve the association of metal
atoms with nitrogen moieties immobilized in a conductive carbon matrix. Recent research has
lead to increased understanding of key aspects of catalyst activation, function, and stability.
However, the nature of the active site is not yet clear because of the complexities introduced by
pyrolysis on high surface-area supports. Such lack of understanding has hindered the engineering
and implementation of these catalysts in fuel cell applications.

28#

Dodelet et al. suggest that MNC catalytic sites occur where metal atoms are bridge-bonded to
nitrogen sites and span micropores of width less than 2 nm.4,5 Lefevre et al. also hypothesize that
high temperature ammonia etching of micropores during pyrolysis creates potential catalytic site
hosts.6 Although microporosity can have a significant impact on activity, species transport to
reactive sites occurs most efficiently through mesopores. This suggests that, for geometrically
complex materials, mesoporosity will have a significant impact on the apparent catalyst
activity.7-9
Although a number of MNC catalyst synthesis methods have been reported, including
ammonia etching 10 and ion sputtering,11 this work considers a high-pressure, closed-vessel
pyrolysis.1,12 Precursors are pre-mixed and vacuum-sealed in a quartz vial so that the process
pressure is elevated by evaporation and decomposition of volatile precursors.1 Precursor
retention in the closed vessel and high partial pressures of reaction intermediates leads to
increased catalyst activity.
The present work considers the impact of varying the carbon black substrate on resulting
nitrogen and iron content, and catalyst activity. Melamine and Iron acetate were selected as
nitrogen and iron precursors respectively consistent with our previous work,12 in which high
nitrogen to carbon ratios are shown to improve activity. Nitrogen and iron content were also kept
the same: 6.3 wt% and 1 wt% respectively. This composition was previously found to lead to
optimum ORR current density (data not shown). The carbons used are Ketjen EC 600J, Black
Pearls 2000, Norit SX Ultra, and Vulcan XC 72R. These carbons were selected to examine a
wide variety of pore distributions, from Vulcan with little porosity of any size, to Black Pearls
with substantial mesoporosity (2-50 nm scale pores) and microporosity (<2 nm pores). Norit is a
primarily microporous activated carbon, and Ketjen is a primary mesoporous carbon black.4,6,13

29#

A number of works have explored porosity in MNC catalysts.6,14-16 These articles have
primarily observed the importance of microporosity on catalyst synthesis and etching by nitrogen
precursors, particularly ammonia. To add to the previous body of work, the current study
evaluates the impact of mesoporosity on catalyst properties. A correlation between mesoporosity,
nitrogen content, and activity was identified, indicating the importance of mesoporosity for
transport of reactants during both active site synthesis and ORR catalysis.
2.3 Experimental
2.3.1

Materials

Iron (II)-acetate, melamine, Nafion® solution (5 wt%) and sulfuric acid (ACS grade) were
obtained from Alfa Aesar (Ward Hill, MA). Ketjenblack® 600JD carbon black was obtained
from Akzo Nobel (Chicago, IL). Black Pearls 2000 and Vulcan XC72R carbon blacks were
obtained from Cabot Corporation (Boston, MA). Norit SX Ultra activated carbon was obtained
from Norit Americas (Marshall, TX). Pressurized oxygen cylinders were obtained from Airgas
(Lansing, MI). All materials were used as received. 190 Proof Ethanol was obtained from Koptec
(King of Prussia, PA)
2.3.2

Catalyst Synthesis

7.5 mg Melamine, 31 mg iron acetate, and 850 mg of various carbons were dispersed in
ethanol, which was evaporated under continuous stirring to yield a dry powder. Melamine was
added to the powder to obtain a mass ratio of 6.3 wt% Nitrogen, and the mixture was inserted
into a quartz ampule that was then flame sealed under vacuum. The ampule was subjected to heat
treatment at 800°C for three hours. After pyrolysis, the quartz ampule was broken under
ventilation, releasing a quantity of gas with an ammonia odor. The remaining powder was then
poured out of the opened ampule.

30#

2.3.3

Physical Characterization

Surface area and pore size distribution were obtained using a Micromeretics ASAP 2020
Surface Area and Porosity Analyzer. Nitrogen adsorption isotherms at 77 K were measured. The
Brunauer, Emmett, and Teller (BET)17 surface area calculation, Barrett, Joyner, and Halenda
(BJH)18 pore volume calculation and non-local density functional theory (NLDFT) were used to
analyze the isotherms. The NLDFT model is a non-graphitic model developed by Ustinov.19,20
Catalyst nitrogen content was evaluated by carbon hydrogen nitrogen (CHN) analysis (Midwest
Microlab, Indianapolis, IN). Iron content was measured using the inductively coupled plasma
(ICP) technique (Michigan State University Department of Geological Sciences, MI). X-ray
diffraction (XRD) spectra of MNC catalysts obtained using a Bruker AXS Model D8 Advance
diffractometer with Cu Kα radiation (λ = 1.54 Å). Thermogravimetric analysis (TGA) results
were obtained using a TGA Q50 (TA instruments). Temperature was scanned at 5 °C min-1
below 450 °C and 20 °C min-1 from 450 to 850 °C. Chemical characterization of TGA products
was obtained using a JSM 6610LV Scanning Electron Microscope (SEM) from JEOL with
energy dispersive X-ray spectrometer (EDS) with acceleration voltage set to 8 kV.
2.3.4

Electrode Preparation

Catalyst ink was prepared by dispersing 4 mg of the catalyst powder ultrasonically in a
solution mixture containing 150 µL ethanol and 50 µL Nafion® (5 wt% solution). A catalyst
loading of 0.5 mg cm−2 was achieved by depositing 5 µL of the suspension on the glassy carbon
electrode followed by drying for 10 min at 60°C.
2.3.5

Electrochemical Characterization

Electrochemical characterization was conducted using a glassy carbon rotating ring-disk
electrode (RRDE, 5 mm diameter, Pine Research Instrumentation, Raleigh, NC). All experiments

31#

were conducted in O2-saturated 0.5 M sulfuric acid (pH 1) at 40 °C using a K2SO4 saturated
Hg|Hg2SO4 reference electrode (0.65 V/NHE); all potentials were corrected to a reversible
hydrogen electrode (RHE) reference based on a measured electrolyte pH of 0.8. A platinum wire
coil served as the counter electrode. Linear-sweep polarization curves were recorded in the
potential range 1.0–0.4 V/RHE at a scan rate of 0.5 mV s−1 and rotation speed of 1200 rpm. In
RRDE experiments, the ring was poised at 1.2 V/RHE. Peroxide percentage, XH2O2, was
calculated from the collection efficiency (N = 0.24), ring current, Iring, and disk current, Idisk

Χ H 2O2 =

2I ring / N
*100%
I disk + I ring / N

Ring collection efficiency represents the fraction of reaction products produced at the disk that
are collected at the ring.1 This collection efficiency was validated by reduction of ferricyanide at
Pt disk and re-oxidation at the ring as explained in detail by Paulus et al.21 Further experiments
confirmed that a MNC catalyst deposition on the Pt disk had no significant impact on collection
efficiency.
2.4 Results and Discussion
In this work iron acetate and melamine are used as precursors on various carbon substrates.
Pore size distributions obtained by NLDFT of nitrogen adsorption data are shown in Figure 2.1.
A wide variety of pore distributions are observed, from Vulcan with little porosity of any size, to
Black Pearls with substantial mesoporosity (2-50 nm scale pores) and microporosity (<2 nm
pores). The catalyst morphologies are similar to the neat carbon substrates with the main
differences in the microporous region. Most of the catalysts lose pore volume during pyrolysis.
Norit loses no porosity (Fig 2.1c), while Black Pearls (Fig 2.1b), and Ketjen (Fig 2.1a), lose
substantial microporosity. Vulcan has very little porosity to begin with. A potential explanation
for decreased catalyst microporosity is deposits from decomposition products, primarily from the
32#

melamine precursor.1,12 In the case of Norit, the lack of mesoporosity may prevent
decomposition products from collecting in the micropores, allowing etching from high
temperature nitrogen intermediates such as ammonia. Numerical values for the surface area and
pore volume are included in Table 2.1.

Table 2.1: Physical and Electrochemical Characteristics of MNC Catalysts

Carbon
Substrate Material

Surface Area
(m2 g-1 by BET)
Support

Catalyst

Catalyst Pore Volume
(cm3 g-1 by DFT)
Micropore
(<2 nm)
±13%

Mesopore
(2-50 nm)
±8%

N content
(wt%)

Activity
mA/mg
@0.8 V/RHE

Ketjen EC600J

1439

1021

0.39

1.28

4.2±0.4

2.23±0.67

Black Pearls 2000

1426

688

0.35

0.55

3.4±0.4

0.13±0.09

Norit SX Ultra

938

993

0.19

0.23

0.4±0.4

0.06±0.01

Vulcan XC 72R

269

194

0.038

0.19

0.8±0.4

0.02±0.01

33#

Figure 2.1: Pore Size Distribution of various carbon precursors (open shapes) and the
associated catalysts (filled shapes) modeled from DFT calculations based on nitrogen
adsorption at 77K. a) Ketjen, b) Black Pearls, c) Norit, d) Vulcan.

Rotating ring-disc electrode (RRDE) studies were performed on the four catalysts. Typical
linear sweep polarization curves are shown in Figure 2.2. Catalysts using Ketjen show not only a
higher activity than those from other substrates, they also show lower peroxide generation of
about 10 percent at 0.8 V vs. RHE decreasing to under 5 percent below 0.5 V vs. RHE. The other
catalysts had low activity around 0.8 V vs. RHE preventing an accurate measurement of peroxide
generation at that potential. Peroxide generation at 0.5 V vs. RHE for Norit, Black Pearls, and
Ketjen were 10, 15, and 20 percent respectively. Except for Norit, the relative peroxide
generation correlates with activity. Ketjen-based catalysts generate the least peroxide, followed
by Norit, Black Pearls, and Vulcan. Because high peroxide generation implies low current
efficiency—i.e. not all the oxygen is fully reduced to water—the low peroxide generation of the
34#

relatively inactive Norit catalyst suggest that it may have good catalytic activity for peroxide
reduction, which prevents peroxide from escaping the catalyst layer.

Figure 2.2: Oxygen reduction at rotating ring-disk electrodes (RRDE) with percent
peroxide generation. Experimental conditions: O2-saturated 0.5 M H2SO4, 60°C,
rotation speed 1200 rpm, catalyst loading of 500 µg/cm 2 on a glassy carbon electrode.

The Tafel plots shown in Figure 2.3 represent mass transfer corrected RDE data, where the
key metric is current density at 0.8 V/RHE values of which are listed in Table 3. Mass transfer
corrections were accomplished using a Koutecky-Levich relationship:

1 1 1
= +
i ik i pl
Were the measured current, i, is related to a kinetically limited current ik and the plateau limiting
current, ipl, which was measured from the plateau of the polarization curve. In this way we were
able to isolate and report the kinetically limited current for any given potential. As shown in
35#

Figure 2.3, all catalysts exhibit similar Tafel slopes of ~80 mV/decade transitioning to a second
larger slope below 0.7 V/RHE. The smaller initial slopes of 80 mV/decade slope are similar to
other results for MNC catalysts,1,22,23 Perry et al. suggest that Tafel slopes in porous electrodes
can be controlled by electrochemical kinetics, mass transport, and ohmic resistance.24 The
similarity in Tafel slopes suggests that the ORR mechanism does not vary significantly among
these catalysts.

Figure 2.3: Tafel plot of oxygen reduction polarization of four different catalysts.
Conditions as in Fig. 2.

A correlation was observed between mesoporous volume and activity as shown in Figure 2.4.
Nitrogen content of the catalysts, as determined by CHN analysis, also scales with the
mesoporosity and electrochemical activity. Although no correlation was found between virgin
microporosity and activity, the correlation indicated by Figure 2.4 suggests that, for fabrication
36#

of catalysts by high-pressure pyrolysis, activity is heavily impacted by a combination of transport
effects caused by mesoporosity and the ability of substrates to adsorb nitrogen precursors.

Figure 2.4: Relationship between mesoporous volume of four different catalysts and
two different parameters: nitrogen content measured by CHN, and iR-free oxygen
reduction current density at 0.8 V vs. RHE.

A second correlation was observed between nitrogen content and iron content as shown in
Figure 2.5. These catalysts were analyzed using CHN combustion analysis to measure bulk
nitrogen content, and by inductively coupled plasma (ICP) to assess iron content. Carbon
materials that retained high iron content after pyrolysis also retained high nitrogen contents.
Although nominal iron and nitrogen content were fixed at 4.4 at% and 6.7 at%, respectively, for
all catalysts, the final products had varying retention of these initial loadings, and therefore
various concentrations of both.

37#

Figure 2.5: Correlation between nitrogen content and iron content in catalysts prepared
with various carbon precursors.

Nitrogen is likely lost via ammonia evolution, as evidenced by the strong smell upon opening
the pyrolyzed ampule. The mechanism of iron loss is not so obvious. Iron may be lost during the
mixing process in ethanol, due to varying levels of iron adsorption on the various carbons.
Unadsorbed iron acetate in suspension adheres to the container walls upon drying. Some ironcontaining compounds, such as iron pentacarbonyl have high vapor pressures at room
temperature.25 These compounds could evaporate once the pyrolyzed ampule is opened.

38#

Figure 2.6: TGA of various precursor materials. Mixture represents a 6.3 to 1 nitrogen
to iron mixture of melamine and iron (II) acetate.

Table 2.2: Composition of TGA Product by EDS*
Material

Fe
(wt%)

C
(wt%)

O
(wt%)

Melamine, Iron (II)
Acetate Mixture
(Fe:N::1:6.3)

58±13

26±7

16±7

*No significant nitrogen content
In order to study iron loss via volatile compounds, TGA was performed on iron acetate,
melamine, and an iron-melamine mixture in an argon atmosphere. Figure 2.6 shows a plot of
sample mass remaining, normalized to original iron content of the precursors (left vertical axis),
as a function of temperature. Melamine samples fully volatilize above 300C, resulting in
complete mass loss. Iron acetate also loses mass associated with the acetate moieties, and only
iron carbide or oxide remains above 300C. One can observe that, for the mixture of iron (II)
39#

acetate and melamine, the relative mass falls well below one above 600 C, suggesting iron loss to
the gas phase. EDS of the post-TGA product indicated that the sample contained only about 58
wt% iron (shown in Table 2.2), meaning that only about half of the original iron is retained after
TGA. Iron may therefore be lost via a volatile iron compound at high temperatures. The
detailed mechanism of this iron loss is a subject of future research.
As evident in Figure 2.5, high iron content correlates to high nitrogen content, which
suggests that iron content may encourage nitrogen adsorption, or vice versa. When plotted on a
molar basis, as at%, the slope of the trend-line is 3, suggesting a 3:1 atomic ratio of nitrogen to
iron in these catalysts. This result is in the right range for most model active sites, which
consider either 2:1 or 4:1 ratios.26,27 The x-intercept of the plot is non-zero (2.8 at%), which
suggests that a large portion of the iron content is not involved in nitrogen adsorption. Such iron
may exist as bulk metal nanoparticles with limited contact with carbon and nitrogen species.
The correlation of nitrogen content with porosity is weak, as shown in Figure 2.4 (right axis),
mainly because the activated carbon Norit displayed particularly weak nitrogen absorption. It is
likely that chemical or surface properties also contribute to nitrogen adsorption. This would
suggest that the various carbon blacks have different optimal nitrogen contents. Various nitrogen
contents have been studied on these carbon blacks (data not shown), but the results did not
change observed relative activity. For that reason XRD data was collected giving a qualitative
assessment of catalyst graphiticity as shown in Figure 2.7. While the majority of the catalysts
have the graphitic d-peak around 24 degrees, Norit’s peak is shifted about 2 degrees. Although
such values of d-peak are too far from graphitic carbon to be quantitatively meaningful, the
larger apparent d-spacing (smaller d-peak angle) in Norit signifies a lower degree of graphiticity
as compared to the other carbons, which could have a significant impact on iron and/or nitrogen

40#

adsorption.

Figure 2.7. X-ray diffraction (XRD) spectra of catalysts prepared with various carbon
precursors. Curves offset by 1000 au.

2.5 Conclusions
An electrochemical and morphological study of various carbon supports in MNC
electrocatalysts for oxygen reduction indicates that carbon mesoporosity can impact nitrogen
adsorption in the catalyst and overall electrochemical activity. The relationship between nitrogen
adsorption and activity shows that substrates that adsorb the most nitrogen and iron show the
highest activity. Moreover, ICP results show that interactions between the carbon support and
iron content also play a role in nitrogen adsorption. The relationship between mesoporosity and
nitrogen adsorption may indicate the importance of transport of precursors to potential active
sites. As mesoporosity increases, both nitrogen adsorption and activity increase. This relationship

41#

is likely influenced by variations in surface chemistry and microstructure.
2.6 Acknowledgments
The authors gratefully acknowledge support from the United States Dept. of Energy Hydrogen
and Fuel Cell Program via Northeastern University. We are also grateful to Dr. Eugene Ustinov
for assistance with NLDFT analysis.
This work is based on results previously published in ECS Transactions 4:1 pp. 1175 (2011).

42#

REFERENCES

43#

REFERENCES

1.

R. Kothandaraman, V. Nallathambi, K. Artyushkova and S. C. Barton, "Non-precious
oxygen reduction catalysts prepared by high-pressure pyrolysis for low-temperature fuel
cells," Applied Catalysis B Environmental, 92(1-2), 209-216 (2009).
doi:10.1016/J.Apcatb.2009.07.005

2.

M. Bron, S. Fiechter, M. Hilgendorff and P. Bogdanoff, "Catalysts for oxygen reduction
from heat-treated carbon-supported iron phenantroline complexes," Journal of Applied
Electrochemistry, 32(2), 211-216 (2002). doi:10.1023/a:1014753613345

3.

H. T. Chung, C. M. Johnston, K. Artyushkova, M. Ferrandon, D. J. Myers and P. Zelenay,
"Cyanamide-derived non-precious metal catalyst for oxygen reduction," Electrochemistry
Communications, 12(12), 1792-1795 (2010). doi:10.1016/j.elecom.2010.10.027

4.

F. Jaouen, S. Marcotte, J. P. Dodelet and G. Lindbergh, "Oxygen reduction catalysts for
polymer electrolyte fuel cells from the pyrolysis of iron acetate adsorbed on various
carbon supports," Journal of Physical Chemistry B, 107(6), 1376-1386 (2003).
doi:10.1021/Jp021634q

5.

M. Lefevre, E. Proietti, F. Jaouen and J. P. Dodelet, "Iron-Based Catalysts with Improved
Oxygen Reduction Activity in Polymer Electrolyte Fuel Cells," Science, 324(5923), 7174 (2009). doi:10.1126/science.1170051

6.

M. Lefevre and J. P. Dodelet, "Fe-based electrocatalysts made with microporous pristine
carbon black supports for the reduction of oxygen in PEM fuel cells," Electrochimica
Acta, 53(28), 8269-8276 (2008). doi:10.1016/j.electacta.2008.06.050

7.

S. H. Joo, S. J. Choi, I. Oh, J. Kwak, Z. Liu, O. Terasaki and R. Ryoo, "Ordered
nanoporous arrays of carbon supporting high dispersions of platinum nanoparticles,"
Nature, 412(6843), 169-72 (2001). doi:10.1038/35084046

8.

S. Pylypenko, T. S. Olson, N. J. Carroll, D. N. Petsev and P. Atanassov, "Templated
Platinum/Carbon Oxygen Reduction Fuel Cell Electrocatalysts," Journal of Physical
Chemistry C, 114(9), 4200-4207 (2010).

9.

E. Proietti, F. Jaouen, M. Lefevre, N. Larouche, J. Tian, J. Herranz and J. P. Dodelet,
"Iron-based cathode catalyst with enhanced power density in polymer electrolyte
membrane fuel cells," Nature communications, 2, 416 (2011). doi:10.1038/ncomms1427

10.

M. Lefevre, J. P. Dodelet and P. Bertrand, "O2 Reduction in PEM Fuel Cells: Activity
and Active Site Structural Information for Catalysts Obtained by the Pyrolysis at High
Temperature of Fe Precursors," The Journal of Physical Chemistry B, 104(47), 1123811247 (2000). doi:10.1021/jp002444n

44#

11.

R. Yang, A. Bonakdarpour, E. B. Easton, P. Stoffyn-Egli and J. R. Dahn, "Co--C--N
Oxygen Reduction Catalysts Prepared by Combinatorial Magnetron Sputter Deposition,"
Journal of The Electrochemical Society, 154(4), A275-A282 (2007).

12.

V. Nallathambi, N. Leonard, R. Kothandaraman and S. Calabrese Barton, "Nitrogen
Precursor Effects in Iron-Nitrogen-Carbon Oxygen Reduction Catalysts,"
Electrochemical and Solid State Letters, 14(6), B55-B58 (2011). doi:10.1149/1.3566065

13.

S. Vengatesan, H.-J. Kim, S.-K. Kim, I.-H. Oh, S.-Y. Lee, E. Cho, H. Y. Ha and T.-H.
Lim, "High dispersion platinum catalyst using mesoporous carbon support for fuel cells,"
Electrochimica Acta, 54(2), 856-861 (2008). doi:10.1016/j.electacta.2008.05.072

14.

F. Jaouen, M. Lefevre, J. P. Dodelet and M. Cai, "Heat-treated Fe/N/C catalysts for O-2
electroreduction: Are active sites hosted in micropores?," Journal of Physical Chemistry
B, 110(11), 5553-5558 (2006).

15.

S. Barazzouk, M. Lefevre and J. P. Dodelet, "Oxygen Reduction in PEM Fuel Cells: FeBased Electrocatalysts Made with High Surface Area Activated Carbon Supports,"
Journal of the Electrochemical Society, 156(12), B1466-B1474 (2009).
doi:10.1149/1.3242293

16.

F. Jaouen, J. Herranz, M. Lefevre, J.-P. Dodelet, U. I. Kramm, I. Herrmann, P. Bogdanoff,
J. Maruyama, T. Nagaoka, A. Garsuch, J. R. Dahn, T. Olson, S. Pylypenko, P. Atanassov
and E. A. Ustinov, "Cross-Laboratory Experimental Study of Non-Noble-Metal
Electrocatalysts for the Oxygen Reduction Reaction," ACS Applied Materials &
Interfaces, 1(8), 1623-1639 (2009). doi:10.1021/am900219g

17.

S. Brunauer, P. H. Emmett and E. Teller, "Adsorption of gases in multimolecular layers,"
Journal of the American Chemical Society, 60, 309-319 (1938).

18.

E. P. Barrett, L. G. Joyner and P. P. Halenda, "The Determination of Pore Volume and
Area Distributions in Porous Substances .1. Computations from Nitrogen Isotherms,"
Journal of the American Chemical Society, 73(1), 373-380 (1951).

19.

E. A. Ustinov, D. D. Do and V. B. Fenelonov, "Pore size distribution analysis of
activated carbons: Application of density functional theory using nongraphitized carbon
black as a reference system," Carbon, 44(4), 653-663 (2006).
doi:10.1016/j.carbon.2005.09.023

20.

E. A. Ustinov, "Nitrogen Adsorption on Silica Surfaces of Nonporous and Mesoporous
Materials," Langmuir, 24(13), 6668-6675 (2008). doi:10.1021/la704011z

21.

U. A. Paulus, T. J. Schmidt, H. A. Gasteiger and R. J. Behm, "Oxygen reduction on a
high-surface area Pt/Vulcan carbon catalyst: a thin-film rotating ring-disk electrode
study," Journal of Electroanalytical Chemistry, 495(2), 134-145 (2001).

45#

22.

T. E. Wood, Z. Tan, A. K. Schmoeckel, D. O'Neill and R. Atanasoski, "Non-precious
metal oxygen reduction catalyst for PEM fuel cells based on nitroaniline precursor," in
510-516, (Elsevier Science Bv, 2008). doi:10.1016/j.jpowsour.2007.10.093

23.

A. A. Gewirth and M. S. Thorum, "Electroreduction of Dioxygen for Fuel-Cell
Applications: Materials and Challenges," Inorganic Chemistry, 49(8), 3557-3566 (2010).
doi:10.1021/ic9022486

24.

M. L. Perry, J. Newman and E. J. Cairns, "Mass transport in gas-diffusion electrodes: A
diagnostic tool for fuel-cell cathodes," Journal of the Electrochemical Society, 145(1), 515 (1998).

25.

A. G. Gilbert and K. G. P. Sulzmann, "The Vapor Pressure of Iron Pentacarbonyl,"
Journal of the Electrochemical Society, 121(6), 832-834 (1974). doi:10.1149/1.2401930

26.

J. Masa, T. Schilling, M. Bron and W. Schuhmann, "Electrochemical synthesis of metalpolypyrrole composites and their activation for electrocatalytic reduction of oxygen by
thermal treatment," Electrochimica Acta, 60(0), 410-418 (2012).
doi:10.1016/j.electacta.2011.11.076

27.

M. Ferrandon, A. J. Kropf, D. J. Myers, K. Artyushkova, U. Kramm, P. Bogdanoff, G.
Wu, C. M. Johnston and P. Zelenay, "Multitechnique Characterization of a Polyaniline-Iron--Carbon Oxygen Reduction Catalyst," The Journal of Physical Chemistry C,
116(30), 16001-16013 (2012).

46#

Chapter 3. Solid Alkalinity of MNC Catalyst for Oxygen Reduction Reaction
3.1

Abstract
Low-cost oxygen reduction catalysts composed of pyrolyzed metal/nitrogen/carbon (MNC)

compounds involve the association of metal atoms with nitrogen species integrated into a
conductive carbon support. Currently, the nature of the active site is not completely understood
due to complexities introduced by pyrolysis on high surface-area supports. The present work
considers the impact of surface alkalinity on the activity of MNC catalysts prepared using a highpressure pyrolysis method. The basic site strength and quantity are calculated by two different
methods, and it is shown that increased Brønsted-Lowry basicity correlates to more active
catalysts.
3.2

Introduction
The cost of precious metal oxygen reduction catalysts represents a major barrier to

commercialization of low temperature fuel cells. Low-cost oxygen reduction catalysts composed
of pyrolyzed metal/nitrogen/carbon (MNC) compounds1,2 have been studied for over 40 years
but have recently been demonstrated to possess oxygen reduction activity nearly high enough to
replace platinum-based catalysts. In the literature the impact of porosity, nitrogen content, and
many other parameters on catalyst activity have been explored, but the nature of the active site is
not completely understood due to complexities introduced by pyrolysis on high surface-area
supports. Further understanding of the active site will allow better optimization of the catalysts,
and lead to increased understanding of oxygen reduction.
The present work considers the impact of surface basicity on the activity of MNC catalysts.
Pyrolysis is conducted in a closed vessel containing metal, nitrogen, and carbon precursors,

47!

!

thereby controlling mass losses due to convection of high temperature byproducts. The carbons
used are Ketjen EC 600J, Black Pearls 2000, and Vulcan XC 72R.
Previous reports have considered many important elements of NPM catalysts such as
nitrogen content,1,3 nitrogen precursors,2-4 iron precursors,5 carbon precursors,3,6,7 and heat
treatment 1. Previous work has highlighted the importance of surface chemistry to identification
and understanding of the active site(s).6 The present work elucidates one aspect of surface
chemistry for MNC catalysts, Brønsted basicity, and how that aspect could impact catalyst
function.
Studies by Herranz et al. have highlighted the importance of surface chemistry to
identification and understanding of the active site(s),6 and they have concluded that the turnover
frequency of MNC catalysts is dependent on basic nitrogen groups.8 Kramm et al. have
suggested that the principal active site is an iron-nitrogen complex with a basic neighbor nitrogen
functionality that changes the energy scheme of the central iron atom.8 Both researchers suggest
that alkalinity would have a significant impact on catalyst activity based on correlations between
surface chemistry and activity, but the causality is not clear. For this reason the proposed work
attempts to increase the understanding of alkalinity in APMNC catalysts.
3.3

Experimental Methods

3.3.1

Materials

Iron(II)-acetate, melamine, Nafion® solution (5 wt%) and sulfuric acid (ACS grade) were
obtained from Alfa Aesar (Ward Hill, MA). Ketjenblack® 600JD carbon black was obtained
from Akzo Nobel (Chicago, IL). Black Pearls 2000 and Vulcan XC72R carbon blacks were
obtained from Cabot Corporation (Boston, MA). Pressurized oxygen cylinders were obtained
from Airgas (Lansing, MI). All materials were used as received.

48!

!

3.3.2

Catalyst Synthesis

All catalysts were prepared as described previously.6 Briefly, melamine, iron acetate, and
carbon materials were dispersed in 190 proof ethanol, which was then evaporated to yield a dry
powder. After addition melamine at 50 wt% of the dry powder, the sample was flame-sealed in a
quartz ampule under vacuum, and subjected to heat treatment at 800°C for three hours. The
enclosed ampule provides retention of volatile precursors as well as increased activity due to
elevated pressure.
3.3.3

Alkalinity Characterization

The surface pH of the neat carbon supports and the as-prepared catalysts was determined by
sonicating suspensions of 0.5-8 g/L solids in a 10 mL, 0.1 M aqueous KCl solution. After a five
minute resting period, The pH was of this suspension was measured (Fischer Scientific Accumet
AB15) relative to the pure KCl solution (pH 6.5±0.5), representing the impact of the solid on the
electrolyte.9
3.3.4

Electrode Preparation

The catalyst ink was prepared as previously described,6 Briefly, 4 mg of catalyst was
ultrasonically dispersed in a mixture of 150 µL of ethanol and 50 µL of 5 wt% Nafion® solution.
Rotating disk electrodes were prepared at a catalyst loading of 0.5 mg cm−2 by depositing an
appropriate amount of the suspension on a glassy carbon rotating disk electrode (RDE, 5 mm
diameter, Pine Research Instrumentation, Raleigh, NC) and drying for 10 min in air.
3.3.5

Electrochemical Characterization

All experiments were conducted in O2-saturated 0.5 M sulfuric acid at room temperature
using a K2SO4 saturated Hg|Hg2SO4 reference electrode (0.65 V/NHE); all potentials were
corrected to the RHE scale. A platinum wire served as the counter electrode. Linear-sweep

49!

!

polarization curves were recorded in the potential range 1.0-0.4 V/RHE at a scan rate of
0.5 mV s−2 and rotation speed of 1200 rpm.
3.4

Results and Discussion
Increased pH of the catalyst slurries represent a decrease in protons in the aqueous KCl

caused by retention of basic surface groups on the catalyst surface (schematic in Figure 3.1). In
this way, high pH indicates solid basicity. Considering that the catalyst is known to have a wide
variety of nitrogen groups,1 one would expect pH to vary with nominal nitrogen content due to
the known proton affinity of nitrogen based compounds. Schoone et al use molecular dynamics
to calculate proton affinities for eight different pyridine and imidazole based derivatives,
confirming that these nitrogen containing molecules are basic10 In fact, pH closely followed
catalyst activity and not nominal nitrogen content indicating that solid basicity and catalytic
activity are closely related. This relationship could take a number of different forms. For
example, the increased basicity could be a result of optimized chemisorption of nitrogen onto the
catalyst surface during pyrolysis. Although a complete understanding of this process would
require further testing.

50!

!

Figure 3.1: Schematic of catalyst interaction with an aqueous KCl solution where basic surface groups
retain protons from the electrolyte reducing the pH of the slurry.

A second consideration is the relationship between specific surface area and pH. If solid
basicity were tied solely to specific surface area and not to quantity of active sites, one would
expect an increase in pH with an increase in specific surface area. Previous studies have shown a
decrease in surface area with increasing nitrogen content for catalysts prepared by high-pressure
pyrolysis.1 Nallathambi et al use nitrogen precursors of various nitrogen to carbon ratios to show
that increased carbon in precursors leads to decreased surface area.11 The observed maximum in
catalyst basicity for varying nitrogen content suggests that surface area does not significantly
control surface basicity.

51!

!

Figure 3.2: Boehm Titration: suspension pH as a function of initial solution pH. Experimental conditions:
sonicated 2 gL-1 catalyst slurries; acid solutions obtained by dilution of sulfuric acid, base solutions by
solutions of sodium hydroxide.

In to order to calculate average pKa and quantity of basic sites on the catalyst, a Boehm
titration was performed as explained by Herranz et al.12 In short, aqueous solutions of varying
concentrations of H2SO4 and NaOH are sonicated to provide a spectrum of pH values. Catalyst is
added to the solutions, and the slurries are sonicated. After sonication the pH is measured again.
Figure 3.2 represents a plot of catalyst slurry pH as a function of initial solution pH prior to
catalyst insertion. The data can be modeled with the following equations:

pH i = log[B0 ]

(1)

pH f = 7 + ½(pK a + log[B0 ])
(2)
where pHi is the initial pH prior to addition of solid, B0 is the concentration of protonated surface
sites, pHf is the pH after addiction of solid, and pKa is the acid dissociation constant. The first
equation represents the first inflection (around pHi = 3) caused by the basic material in the
52!

!

presence of a concentration of a strong acid (H2SO4) at which all the basics sites become
protonated and there are essentially no remaining excess protons in the solution. The second
equation represents the plateau caused by the basic material in the presence of an essentially
neutral solution at which the material is modeled as a weak base. The resulting curve was used to
calculate an average pKa (7.3 ± 0.7) and quantity of basic sites (0.6 ± 0.5 mmol g-1). These
values agree well with what Herranz et al. calculated for their ammonia based catalyst,12 and
these results also suggest that the vast majority of basic sites are weak bases because the average
acid dissociation constant is on the same order of magnitude as that of water.
The correlation between activity and basicity was confirmed in an optimization study of
nominal nitrogen content shown in Figure 3.3. Varying amounts of nitrogen precursor were
added to the precursor mixture before pyrolysis. An optimal nitrogen proportion was found at

!

!

Figure 3.3: Comparison of nominal nitrogen content to current and basicity of the catalyst.
2
Experimental conditions: RDE: 0.8 V vs. RHE, 0.5M H2SO4, 1200 rpm, catalyst loading 0.5 mg/cm on
glass carbon electrode, pH: 1 g/L catalyst suspension, 0.1 M KCl.

53!

!

around 22 wt.% N. Catalyst basicity did not vary directly with nitrogen content, as would be
expected if basicity were only caused by nitrogen content in the catalyst. Instead pH closely
followed catalyst activity confirming that the active site is associated with basic sites.
A second consideration should be the relationship between specific surface area and pH. If
catalyst proton adsorption were tied solely to specific surface area and not to quantity of active
sites, one would expect an increase in pH with an increase in specific surface area. Previous
studies have shown an general decrease in surface area with increasing nitrogen content for
catalysts prepared by high-pressure pyrolysis.1 Nallathambi et al use nitrogen precursors of
various nitrogen to carbon ratios to show that increased carbon in precursors leads to decreased
surface area.2 The observed maximum in catalyst basicity for varying nitrogen content suggests
that surface area does not significantly control surface basicity.
3.5

Conclusions

An electrochemical and surface basicity study of various MNC electrocatalysts for oxygen
reduction indicates that catalyst basicity is connected to overall electrochemical activity. A trend
was found between catalyst activity and a pH shift in suspension of the catalyst. Higher pH
values represent an increase in basic proton adsorption sites. The trend between catalyst activity
and a pH suggests that basicity is an important parameter in catalyst site activity and
identification.
!

54!

!

!

!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
REFERENCES

55!

!

REFERENCES
!
!
1.

R. Kothandaraman, V. Nallathambi, K. Artyushkova and S. C. Barton, "Non-precious
oxygen reduction catalysts prepared by high-pressure pyrolysis for low-temperature fuel
cells," Applied Catalysis B Environmental, 92(1-2), 209-216 (2009).
doi:10.1016/J.Apcatb.2009.07.005

2.

V. Nallathambi, N. Leonard, R. Kothandaraman and S. C. Barton, "Nitrogen Precursor
Effects in Iron-Nitrogen-Carbon Oxygen Reduction Catalysts," Electrochemical and
Solid State Letters, 14(6), B55-B58 (2011). doi:10.1149/1.3566065

3.

F. Jaouen, J. Herranz, M. Lefevre, J.-P. Dodelet, U. I. Kramm, I. Herrmann, P.
Bogdanoff, J. Maruyama, T. Nagaoka, A. Garsuch, J. R. Dahn, T. Olson, S. Pylypenko,
P. Atanassov and E. A. Ustinov, "Cross-Laboratory Experimental Study of Non-NobleMetal Electrocatalysts for the Oxygen Reduction Reaction," ACS Applied Materials &
Interfaces, 1(8), 1623-1639 (2009). doi:10.1021/am900219g

4.

J. Tian, L. Birry, F. Jaouen and J. P. Dodelet, "Fe-based catalysts for oxygen reduction in
proton exchange membrane fuel cells with cyanamide as nitrogen precursor and/or porefiller," Electrochimica Acta, 56(9), 3276-3285 (2011).
doi:10.1016/j.electacta.2011.01.029

5.

G. Faubert, G. Lalande, R. Cote, D. Guay, J. P. Dodelet, L. T. Weng, P. Bertrand and G.
Denes, "Heat-treated iron and cobalt tetraphenylporphyrins adsorbed on carbon black:
Physical characterization and catalytic properties of these materials for the reduction of
oxygen in polymer electrolyte fuel cells," Electrochimica Acta, 41(10), 1689-1701
(1996). doi:10.1016/0013-4686(95)00423-8

6.

N. D. Leonard, V. Nallathambi and S. Calabrese Barton, "Carbon Supports for NonPrecious Metal Proton Exchange Membrane Fuel Cells," ECS Transactions, 41(1), 11751181 (2011). doi:10.1149/1.3635650

7.

M. Lefevre and J. P. Dodelet, "Fe-based electrocatalysts made with microporous pristine
carbon black supports for the reduction of oxygen in PEM fuel cells," Electrochimica
Acta, 53(28), 8269-8276 (2008). doi:10.1016/j.electacta.2008.06.050

8.

U. I. Kramm, J. Herranz, N. Larouche, T. M. Arruda, M. Lefevre, F. Jaouen, P.
Bogdanoff, S. Fiechter, I. Abs-Wurmbach, S. Mukerjee and J.-P. Dodelet, "Structure of
the catalytic sites in Fe/N/C-catalysts for O2-reduction in PEM fuel cells," Physical
Chemistry Chemical Physics, 14(33), 11673-11688 (2012). doi:10.1039/c2cp41957b

9.

A. Mukherjee, A. R. Zimmerman and W. Harris, "Surface chemistry variations among a
series of laboratory-produced biochars," Geoderma, 163(3-4), 247-255 (2011).
doi:10.1016/j.geoderma.2011.04.021

56!

!

10.

K. Schoone, J. Smets, R. Ramaekers, L. Houben, L. Adamowicz and G. Maes,
"Correlations between experimental matrix-isolation FT-IR and DFT (B3LYP) calculated
data for isolated 1: 1 H-bonded complexes of water and pyridine or imidazole
derivatives," Journal of molecular structure, 649(1), 61-68 (2003).

11.

V. Nallathambi, N. Leonard, R. Kothandaraman and S. Calabrese Barton, "Nitrogen
Precursor Effects in Iron-Nitrogen-Carbon Oxygen Reduction Catalysts,"
Electrochemical and Solid State Letters, 14(6), B55-B58 (2011). doi:10.1149/1.3566065

12.

J. Herranz, F. Jaouen, M. Lefevre, U. I. Kramm, E. Proietti, J.-P. Dodelet, P. Bogdanoff,
S. Fiechter, I. Abs-Wurmbach, P. Bertrand, T. M. Arruda and S. Mukerjee, "Unveiling NProtonation and Anion-Binding Effects on Fe/N/C Catalysts for O2 Reduction in ProtonExchange-Membrane Fuel Cells," The Journal of Physical Chemistry C, 115(32), 1608716097 (2011). doi:10.1021/jp2042526

!
!

57!

!

Chapter 4. Analysis of Adsorption Effects on a Metal-Nitrogen-Carbon Catalyst using a Rotating
Ring-Disk Study
4.1

Abstract

A steady-state, rotating ring disk study of the oxygen reduction reaction (ORR) was
conducted in acid environment using a pyrolyzed metal/nitrogen/carbon (MNC) electrocatalyst.
Analysis of peroxide generation indicates that ORR proceeds both via a direct four-electron
pathway to water at high potentials and an indirect peroxide pathway at low potentials. Above
0.6 V vs RHE, the direct four-electron pathway to water without a desorbing intermediate
dominates oxygen reduction because peroxide generation is inhibited due to site availability. In
contrast, at potentials below 0.6 V, oxygen reduction begins to shift to the indirect peroxide
pathway due to fast kinetics and higher site availability. The net peroxide generation remains
relatively low over the entire range due to reduction of peroxide to water.
4.2

Introduction

The need for less expensive electrocatalysts for the oxygen reduction reaction (ORR) in low
temperature fuel cells has impeded commercialization of these devices for transportation
applications. Metal/nitrogen/carbon (MNC) compounds represent a major focus in the search for
platinum alternatives.1-3 These compounds involve the combination of metal and nitrogen
components immobilized in a conductive carbon matrix, and are typically pyrolyzed to enhance
activity and stability.4 Although the structure of the MNC active site is poorly understood due to
complexities introduced by pyrolysis on high surface-area supports, an increased understanding
of mechanism provides evidence for possible structures, as well as direction for further
engineering optimization of these catalysts for fuel cell applications. In the present work, a

58

rotating ring-disk electrode analysis is conducted to increase understanding of oxygen and
peroxide reactions occurring on a pyrolyzed MNC catalyst.

Figure 4.1: Reaction schematic showing two pathways: pathway one (Ko,1 and k1) is the complete
reduction of oxygen to water with no desorbing intermediate and pathway two (Ko,2 and k2) is the
two-electron reduction to a desorbing peroxide intermediate that disperses into the catalyst layer
(Kp,2), but also has the potential to be further reduced to water (k3).

The oxygen reduction reaction is considered here to be composed of two pathways (Fig.
4.1).5-10 The first pathway represents direct four-electron reduction to water without a desorbing
intermediate (k1 in Fig. 4.1). The second pathway represents incomplete reduction of oxygen to
hydrogen peroxide (k2), which can desorb from the surface (KP,2),5 disproportionate to water and
oxygen,7 or reduce to water (k3).7 Jaouen found that disproportionation was insignificant for his
Fe-N-C catalyst and that disproportionation reaction rates would have to be orders of magnitude
higher that experimental values to account for the low peroxide yield found in these catalyst.7
We therefore neglect disproportionation in the present work.
Rotating ring-disk electrode (RRDE) experiments are used to distinguish between the above
two pathways. The RRDE allows for controlled, well-characterized transport to and from the
catalyst layer at steady state.11 An analysis can be performed considering the figure of merit,
59

IDN/IR, vs. ω-1/2 where ID is disk current, N is ring electrode collection efficiency, IR is ring
current, and ω is rotation speed.6,8-10 The figure of merit, IDN/IR, represents a ratio of disk current
to flux of intermediates from the disk, and indicates the electronic efficiency of the catalyzed
reaction. This parameter can easily be connected to Koutecky-Levich relationships that are
mathematically developed from kinetic models, from which rate constants may be determined by
comparison to experimental results. Hsueh et al. present five such models for RRDE results,9 and
Anastasijevic et al. present another, more sophisticated, model.6 The Anastasijevic model makes
the distinction of representing the 4-electron pathway not as an ideal direct reduction pathway,
but as a strongly adsorbed pathway.
The difficulty in applying these models to the present system is that they assume a planar
electrode. The present MNC catalysts are porous electrodes with nonzero thickness, adding
transport effects that are not considered in the above models.
Previous RRDE studies of non-precious metal catalysts for oxygen reduction1,12-17 generally
show an increase in peroxide generation by one to two percent per 0.1 mg/cm2 decrease in
loading.1,13,17,18 This suggests that peroxide detection at the ring could be artificially low because
inner layers may further reduce peroxide to water, decreasing peroxide at the ring and increasing
total disk current. Additionally, more active catalysts have been found to produce less peroxide;
this is found both by varying metal content14 and morphology of the carbon support. Such
considerations motivate the need for increased understanding of the ORR mechanism using
MNC catalysts. For the present work we take steady-state measurements in order to insure
uniform transport characteristics as well as minimizing unstable intermediates. This has
significant implications especially due to the impact of transport phenomena on peroxide and
oxygen availability at the electrode surface.

60

In the present work, an analysis of the MNC catalyst is attempted via the figure of merit,
IDN/IR. This method is revised to isolate the indirect ORR pathway and establish a new figure of
merit, CoN/IR. Finally, this new figure of merit is utilized to understand the kinetic and
adsorption limitations of the MNC catalyst in the context of Langmuir Hinshelwood kinetics.
4.3
4.3.1

Experimental Methods
Materials

Carbon, nitrogen, and iron materials used in catalyst synthesis were Ketjen 600JD carbon
black (Akzo Nobel, Chicago, IL), melamine (Alfa Aesar, Ward Hill, MA), and iron (II) acetate
(Alfa Aesar, Ward Hill, MA) respectively. Materials used in characterization were
perfluorosulfonic acid-PTFE copolymer (5 wt%), reagent grade sulfuric acid (Alfa Aesar, Ward
Hill, MA), and oxygen (Airgas, Lansing, MI).
4.3.2

Catalyst Synthesis

Catalysts were synthesized as described previously.1,16,18 Briefly, Ketjen (95.7 wt.%),
melamine (0.8 wt.%), and iron (II) acetate (3.5 wt.%) were dispersed in ethanol and dried
overnight to produce a dry powder. Further melamine was added to the powder in a ratio of 5:9
to increase the nitrogen content to 24 wt.% (resulting iron content was 0.7 wt.%), and 105 mg
were inserted in a quartz ampule. The quartz ampule was pulled under vacuum to a volume of
2.4 cc, and pyrolyzed at 800° C for 3 hours. The quartz ampule was opened and the catalyst
consisted of the contained powder.
4.3.3

Electrode Preparation

The catalyst ink was prepared by ultrasonically mixing 4 mg catalyst in 50 µL of 5 wt.%
ionomer and 150 µL 190 proof ethanol as described previously16, catalyst layers were prepared at

61

loadings of 0.5 to 0.1 mg cm−2 by depositing 5 µL of the ink on the 5 mm diameter, glassycarbon disk of a rotating ring-disk electrode (Pine Research Instrumentation, Raleigh, NC) and
drying for 10 min in air. The ink was diluted with ethanol to achieve lower loadings; in this way
the drop volume and ionomer to catalyst ratio were held constant for all loadings.
4.3.4

Electrochemical Characterization

Experiments were conducted in oxygen saturated 0.5 M sulfuric acid at room temperature.
Potentials were measured relative to a Hg/Hg2SO4 reference electrode that was calibrated with
respect to a reversible hydrogen electrode. All potentials were corrected to RHE scale (0.7 V vs.
RHE). A platinum wire served as the counter electrode. Potentiostatic measurements were made
at 400, 900, 1200, and 1600 rpm in 50 mV increments between 0.2 and 0.9 V vs. RHE.
Fractional peroxide yield, χp, was calculated based on the following equation:19
p

=

2IR /N
ID + IR /N

[1]

Where IR is the ring current, N the collection efficiency, and ID the disk current. Collection
efficiency, N, was calculated from similar potentiostatic measurements in 0.1 M NaOH, 0.01 M
K3Fe(CN)6. The collection efficiency was not found to vary significantly with loading, although
it did decrease slightly with rotation speed consistent with Claude et al.20 Collection efficiencies
were 0.24 ± 0.01, 0.23 ± 0.01, 0.23 ± 0.01, and 0.22 ± 0.01 for 400, 900, 1200, and 1600 rpm
respectively.

62

4.4

Results and Discussion

Potentiostatic, steady-state RRDE experiments were conducted at fixed catalyst loading for
four different disk rotation speeds, and at fixed rotation speed for five different loadings. Fig.
4.2a shows polarization disk current and ring current with respect to disk potential at various
rotation speeds. Above 0.7 V, both the disk current, ID, and ring current, IR, exhibit exponential
(Tafel) dependence on potential. The disk limiting current (Ilim, the current plateau below 0.6 V)
increases with increasing rotation speed, indicating mass transfer limitation. The ring current also
increases with rotation speed, but is more weakly correlated than the disk current. This implies
that IDN/IR increases with rotation speed.
Fig. 4.2b shows steady state polarization curves at various catalyst loadings. Loading impacts
disk current at high potentials, where the current falls below Ilim, but impacts ring current much
more significantly, even at low potentials.
Comparisons can be made between the peroxide generation rate and the overall rate of water
production. Fig. 4.2b shows that above 0.7 V vs. RHE, disk current increases significantly with
loading, whereas ring current is relatively constant. The disk current at such high potentials can
therefore be primarily attributed to four-electron reduction because the ring current is 500-fold
smaller and constant over the same loading range. Hence, the peroxide generating ORR pathway
appears to be insignificant at high potentials.

63

Figure 4.2: Steady-state polarization curve showing disk current and ring current: 0.5 M H2SO4, room
-2
temperature, oxygen saturated, ring poised at 1.2 V vs. RHE at (a) 0.5 mg cm , various rotation
speeds (b) 1600 rpm, various loadings.

The ring current, IR, can be thought of in two ways: as a measure of peroxide flux from the
disk and as a measure of peroxide concentration at the disk. In the following results and
discussion both perspectives will be considered.
The exponential increase of ring current with decreasing potential near 0.7 V vs RHE (Fig.
4.2) suggests that peroxide generation is limited by the kinetics of the first 2e– reduction step,
because the reduction of oxygen to hydrogen peroxide coincides with that potential. Decreasing
ring current below the maximum at 0.5-0.6 V suggests either the onset of step 3 (further
reduction of peroxide to water) or competition between the two different oxygen reduction
reactions (steps 1 and 2). A detailed analysis of the 2e– + 2e– reaction pathway (I2 and I3) is
necessary to compare these two phenomena, via the figure of merit, IDN/IR.6,8-10

64

By assuming a reaction mechanism, an expression for IDN/IR can be obtained in terms of
fundamental rate constants plus the rotation rate.5,8-10 Without considering adsorption (Ignoring
Ko,1 and Ko,2), the expression for the mechanism displayed in Fig. 4.1 is:9
ID N
k1
2 + 2k1 /k2
=1+2 +
k3 !
IR
k2
zP

1/2

[2]

Where k1, k2, and k3 are the rate constants for reduction of oxygen to water, reduction of water to
hydrogen peroxide, and reduction of hydrogen peroxide to water, respectively. The parameter
zp = 0.62Dp2/3ν-1/6 is an abbreviation for the group of constants in the Levich equation:5,6,8,9
imt = 0.62nF Cb D2/3 ⌫

1/6

! 1/2

[3]

Where the mass transport limited current at a rotating disk, imt, is a function of n is the number
of electrons per mole oxygen, F is Faraday’s constant, Cb is a bulk concentration, D is
diffusivity, ν is viscosity, and ω is the rotation speed. The parameter z is calculated from a
correlation of the mass transfer limiting current and ω1/2. Based on a bulk oxygen concentration
of 1 mM21 and a viscosity of 0.01 cm2/s,22 the resulting z values indicate diffusivities of (2.17 ±
0.07) ×10-5 cm2/s for oxygen (which is in the range of reported values: 1.87×10-5 to 2.12×10-5
cm2/s)21 and (1.28 ± 0.25) ×10-5 cm2/s for hydrogen peroxide (reported value: 1.71×10-5
cm2/s).23

65

Figure 4.3: (a) Plot of IDN/IR vs rotation speed for two different electrodes at two different potentials,
0.5 M H2SO4, room temperature, oxygen saturated, ring poised at 1.2 V vs. RHE. (b) Slope and
intercept plot at various loadings and potentials: 0.5 M H2SO4, room temperature, oxygen saturated,
ring poised at 1.2 V vs. RHE.

The figure of merit, IDN/IR, for the current data set was regressed with respect to the inverse
square root of the rotation speed, and sample regressions are shown in Fig. 4.3a, with a plot of
slope vs intercept given in Fig. 4.3b. One issue with the application of this analysis is that Eq.18
predicts a positive slope with respect to ω-1/2, but the data clearly show negative slope.
A requirement for the validity of Eq. 2 is first order kinetics.5 In order to test this requirement,
the variation of disk and ring currents with oxygen concentration is explored as shown in Fig.
4.4. Assuming a Nernst boundary layer at the disk surface, Co can be calculated from the bulk
oxygen concentration and the ratio of the total reduction current at a given potential and the mass
transfer limiting current:9

ID + IR /N
)
Ilim

Co = Co,b (1

66

[4]

where Co,b represents oxygen concentration in the bulk electrolyte, Co is dissolved oxygen
concentration at the catalyst surface, and Ilim the diffusion limited current. At high potential (0.7
V vs RHE), where the peroxide reduction rate is small, IR ~ NI2. This correlation with oxygen (I1
~ ID and I2 ~ IR/N) shows that neither reaction pathway appears to show first order kinetics
(I

Co).
If the disk and ring currents were first order, the traces in Fig. 4.4 would be straight lines that

intercept the origin. Because this is not the case, kinetics controlled by surface adsorption were
explored. For single constituent adsorption, the Langmuir adsorption coefficient may be defined
as: 23
Ki,j =

Ci ✓v,j
✓i,j

[5]

where Ki,j is the adsorption coefficient for species i on site j, θi,j is the fraction of sites occupied
by species i, Ci is concentration of species i near the surface, and θv,j is the fraction of vacant
sites. A mass balance on the total number of sites of type j yields:
X
1 = ✓v,j +
✓i,j
i

[6]

Then θv,j can be solved for and substituted into the adsorption equations to obtain the surface
fractions. For a single adsorbent this results in:
✓i,j =

Ci
Ki,j + Ci

If an electrochemical reaction is first order with respect to the surface species, θi,j, reaction
rate laws can be written as a current:

67

[7]

i=

nF kCi
Ki,j + Ci

[8]

where k is a rate constant, and nFk represents a plateau current density when Co ≫ Ko,j, i.e. when
the surface is saturated.

Figure 4.4: Plot of disk and ring current densities vs oxygen concentration (varying rotation speed)
for two different loadings at 0.7 V vs RHE, 0.5 M H 2SO4, room temperature, oxygen saturated, ring
poised at 1.2 V vs. RHE. Dashed lines represent fit of Eq. 8.

Returning to Fig. 4.4, one can see that Eq. 8 fits the experimental data well. The data are
therefore well explained by a site-limited surface reaction, because the current density plateaus
as C0 increases. As mentioned earlier, a high potential was chosen for Fig. 4.4 to minimize rate 3
(peroxide reduction). The small rate three allows comparison between oxygen adsorption on the
direct 4e– and indirect 2e– active sites. The smaller variation with oxygen concentration in the
ring current (representing peroxide generation) as compared to the disk current (dominated by
4e– oxygen reduction) suggests that the 2e– sites are more saturated than the 4e– sites.
68

One can extend the analysis to lower potentials for the disk current, assuming that I1
dominates. Disk current is plotted vs. oxygen concentration at 0.5 V vs RHE in Fig. 4.5; the disk
current displays more linear kinetics, suggesting that the surface concentration is low, Co ≪ Ko,j.

Figure 4.5: Plot of disk current density vs oxygen concentration (varying rotation speed) for two
different loadings at 0.5 V vs RHE, 0.5 M H2SO4, room temperature, oxygen saturated, ring poised at
1.2 V vs. RHE.

The observed adsorption limitation and resulting nonlinear kinetics do not satisfy the
requirements for Eq. 2. For this reason, the RRDE data must be reanalyzed assuming LangmuirHinshelwood kinetics. In this analysis we will assume that the two reduction pathways occur on
different sites. As shown in Fig. 4.1, site one is responsible for direct 4e– reduction of oxygen to
water, while site two is responsible for the indirect, 2e– + 2e– reduction pathway via peroxide.
We will obtain an expression for parameter CoN/IR, where CO is calculated by Eq. 4, and will
assume that the rate peroxide generation by reaction 2 is first order with respect to oxygen
69

concentration. Initially, the method will be established assuming linear kinetics (no adsorption
limitations), and then extended to consider Langmuir adsorption. For the initial analysis, with
linear kinetics (Co ≪ Ko,j) The two-electron oxygen reduction current I2 may be expressed as:
I2 =

nF k2 Co
Ko,2

[9]

where k2 is the associated rate constant, Ko,2 is the adsorption equilibrium coefficient for 2+2
pathway surface sites (site 2), and Co is oxygen concentration given by Eq. 4. The ring current,
IR, is related to the concentration of peroxide, Cp at the disk by a Levich-like transport model:6,810

2/3 1/6

IR = nF N Cp zp ! 1/2 , where zi = 0.62Di

⌫

[10]

where n is the number of electrons per mole oxygen, F is Faraday’s constant, D the diffusivity, ν
the viscosity, and ω the rotation speed. An equation for I3 with respect to ring current can be
developed by assuming that I3, i.e. the further reduction of peroxide to water is first order with
respect to peroxide concentration at the disk:

nF k3
k3 IR ! 1/2
I3 =
Cp =
Kp,2
Kp,2 zp N

[11]

where Kp,2 is the adsorption coefficient of peroxide on site two. Eqns. 9 and 11 provide rate
expressions for each step associated with the two-by-two reaction pathway (I2 and I3) with the
only parameters being three rate constant groups (k2/Ko,2, k3/Kp,2) and mass transport properties.
A charge balance on peroxide generation/consumption relates the ring current to I2 and I3:
I2 =

IR
+ I3
N
70

[12]

Here we have neglected bulk disproportionation of peroxide. Combining equations 12 and 11
allows I2 to be represented as a function of the ring current and some reaction coefficient k3/Kp,2:
I2 =

IR
k3
(
!
N Kp,2 zp

1/2

+ 1)

[13]

Rearranging Eq. 13 and introducing Co via Eq. 9 yields:
Co N
Ko,2 k3
=
!
IR
nF k2 Kp,2 zp

1/2

+

Ko,2
nF k2

[14]

In the absence of adsorption limitations, the term k2/Ko,2 can then be calculated from a regression
of CoN/IR with rotation speed, allowing the calculation of I2 via Eq. 9. Based on the earlier
analysis of Fig. 4, the present data indicate limitations due to oxygen adsorption. Therefore, an
adsorption model must be incorporated into the reaction system.
For oxygen and peroxide, the Langmuir adsorption coefficients, Ko,2 and Kp,2, may be defined
as:24
Ko,2 =

Co ✓v,2
Cp ✓v,2
; Kp,2 =
✓o,2
✓p,2

[15]

where θo,2 and θp,2 are the surface fractions of oxygen and peroxide respectively on the sites
involved in indirect ORR pathway (site 2), θv,2 is the surface availability of this adsorption site,
and Co, Cp are oxygen and peroxide concentrations near the surface. If the reactions two and
three are first order with respect to the surface species (θo,2 and θp,2) the rate laws can be rewritten:
I2 =

nF k2 ✓v,2 Co
Ko,2

71

[16]

I3 =

nF k3 ✓v,2 Cp
Kp,2

[17]

Using these rate laws, Eq. 14 becomes:
Co N
Ko,2 k3
=
!
IR
nF k2 Kp,2 zp

1/2

+

Ko,2
nF k2 ✓v,2

[18]

Assuming that the surface fractions of oxygen and peroxide are small (that is θv,2 is near unity)
the equation reduces to Eq. 14.

Figure 4.6: Correlation of disk current with hydrogen peroxide concentration showing first-order
-2
-1
peroxide reduction reaction: slope is 0.0564 ± 0.0007 mA cm mM . Hydrogen peroxide
concentrations range from 10 µM to 4 mM in 0.5 M H2SO4: room temperature, nitrogen saturated.

Peroxide reduction experiments were performed by running the same RRDE experiment as
above, but in the presence of various concentrations of hydrogen peroxide and saturated with
nitrogen. By plotting the current at a given potential as a function of peroxide concentration, the
72

presence of adsorption limitations associated with peroxide reduction can be studied. As
demonstrated in Fig. 4.6, in the absence of oxygen, the observed peroxide reduction rate is
linearly dependent on bulk hydrogen peroxide concentration at least to 4 mM. This suggests that
Kp,2 is large and adsorbed peroxide does not contribute significantly to surface coverage.
Considering this approximation, surface fractions for site 2 become θv,2 + θo,2 = 1, and Eq. 18
becomes:
Co N
Ko,2 k3
=
!
IR
nF k2 Kp,2 zp

1/2

+

Co + Ko,2
nF k2

[19]

Additionally, at high potentials where I3 is insignificant, Eq. 19 simplifies to

Co N
Co + Ko,2
=
IR
nF k2

[20]

In these cases the figure of merit CoN/IR should scale with Co. Some representative correlations
are shown in Fig. 4.7a. For this correlation the ratio of the intercept to slope indicates an
adsorption coefficient, and the inverse slope represents a rate constant. The resulting reaction
parameters Ko,2 and k2 are shown in Fig. 4.7b as a function of potential. At low loading and high
potentials, Ko,2 does not vary with potential, while k2 varies exponentially with potential. These
characteristics indicate potential independent adsorption and Tafel kinetics. At high loading and
low pressure Ko,2 decreases and k2 becomes less potential dependent, likely caused by the onset
of I3. By fitting rate parameters in the Tafel region, rate parameters can be established. These rate
parameters can be used to calculate surface coverage and I2. Charge balances can be applied to
deduce I1 and I3.

73

2

Figure 4.7: (a) Plot of C oN/IR vs oxygen concentration at four high potentials with 0.3 mg/cm
loading, 0.5 M H2SO4, room temperature, oxygen saturated, ring poised at 1.2 V vs. RHE. (b) Plot of
peroxide generation parameters k2 and Ko,2 at various loadings and potentials. Reaction parameters
-1
-1
are: Ko,2 = 1.7e-4 ± 1.1e-4 M, k2 = 4.3e-8 ± 0.8e-8 mol s mg and the Tafel slope is 114 ± 6 mV per
decade. Conditions: 0.5 M H2SO4, room temperature, oxygen saturated, ring poised at 1.2 V vs.
RHE.

74

Surface coverage at various potentials can be compared as shown in Fig. 4.8. Surface
coverage was calculated via the definition of the adsorption coefficient (Eq. 15) and a mass
balance on adsorption sites (1 = θv,2 + θo,2):
✓o,2 =

Co
Ko,2
; ✓v,2 =
Co + Ko,2
Co + Ko,2

[21]

As shown in Fig. 4.8, For 0.1 mg cm-2 loading, oxygen surface coverage remains above 70%
down to 0.6 V vs RHE. Low surface coverage at 0.5 mg cm-2 indicates that the high loadings
approach mass-transfer-limiting current faster than lower loadings. This is an indication of high
catalyst utilization.

Figure 4.8. Fractional surface coverage at two different loadings and at various rotation speeds.
Open symbols represent vacant sites, while filled symbols represent occupied sites.

75

The currents for each reaction step can also be compared. Fig. 4.9 shows currents I1, I2, and I3
cumulatively as compared to disk current. It is evident that at high potentials I1 dominates, but I2
becomes significant below 0.7 V and I3 below 0.6 V. The fact that I1 dominates over most of the
potential range justifies the use of disk current as a proxy for I1, as was assumed in the context of
Fig. 4.4.
The adsorption limitation may be why I2 is not significant at higher potentials: at high
potential, oxygen surface concentration is high and available sites for adsorption limit the 2+2
reaction pathway. At low potential oxygen concentration and surface coverage are mass transfer
limited, decreasing adsorption limitations. At these low potentials the 2+2 reaction pathway can
compete better with the direct 4-electron pathway.

Figure 4.9: Polarization curve showing contributions of I1, I2, and I3 with experimental ID.values for
various loadings at 1600 RPM.

76

4.5

Conclusions

An RRDE study indicates that oxygen reduction on this autogenic-pressure MNC catalyst
proceeds primarily through the direct four-electron pathway to water. Langmuir-Hinschelwood
analysis suggests that the indirect 2+2 pathway active sites are more saturated with surface
species.
A second conclusion is that 4e– and 2e– oxygen reduction occur at different sites that compete
for oxygen in the catalyst layer. This is illustrated in the difference in adsorption behaviors
between the disk current and ring current. The variance of disk and ring current with oxygen, as
discussed with relation to Fig. 4.4, indicates that the 2e– adsorption site is more saturated than the
4e– site. Furthermore, correlations between peroxide generation and oxygen concentration
indicate that peroxide generation is significantly adsorption limited.
Together, these main conclusions show a more complete picture of oxygen reduction and
peroxide generation on MNC catalysts: the onset of peroxide generation occurs around its redox
potential of 0.8 V vs RHE with fast kinetics driving a sharp increase in peroxide generation.
Although the peroxide generation does not require a large over-potential, low site availability
keeps the ring current small. As the potential decreases further, the combination of increasing
over-potential and increasing peroxide surface concentration causes a dramatic increase in ring
current. Below 0.6 V a second reduction step scavenges the peroxide on the catalyst surface
decreasing the flux of peroxide from the catalyst layer.
4.6

Acknowledgments

We gratefully acknowledge the partial financial support from the U.S. Department of Energy
(EERE), under a Non PGM Catalyst development effort (Contract no EE 0000459) lead by
Northeastern University (Sanjeev Mukerjee, P.I.).

77

REFERENCES

78

REFERENCES

1.

R. Kothandaraman, V. Nallathambi, K. Artyushkova and S. C. Barton, "Non-precious
oxygen reduction catalysts prepared by high-pressure pyrolysis for low-temperature fuel
cells," Applied Catalysis B Environmental, 92(1-2), 209-216 (2009).
doi:10.1016/J.Apcatb.2009.07.005

2.

M. Bron, S. Fiechter, M. Hilgendorff and P. Bogdanoff, "Catalysts for oxygen reduction
from heat-treated carbon-supported iron phenantroline complexes," Journal of Applied
Electrochemistry, 32(2), 211-216 (2002). doi:10.1023/a:1014753613345

3.

H. T. Chung, C. M. Johnston, K. Artyushkova, M. Ferrandon, D. J. Myers and P. Zelenay,
"Cyanamide-derived non-precious metal catalyst for oxygen reduction," Electrochemistry
Communications, 12(12), 1792-1795 (2010). doi:10.1016/j.elecom.2010.10.027

4.

E. Yeager, "Electrocatalysts for O-2 Reduction," Electrochimica Acta, 29(11), 1527-1537
(1984).

5.

N. A. Anastasijevic, V. Vesovic and R. R. Adzic, "Determination of the kinetic parameters
of the oxygen reduction reaction using the rotating ring-disk electrode: Part I. Theory,"
Journal of Electroanalytical Chemistry and Interfacial Electrochemistry, 229(12), 305-316
(1987). doi:10.1016/0022-0728(87)85148-3

6.

N. Anastasijević, V. Vesović and R. Adžić, "Determination of the kinetic parameters of the
oxygen reduction reaction using the rotating ring-disk electrode: Part II. Applications,"
Journal of electroanalytical chemistry and interfacial electrochemistry, 229(1), 317-325
(1987).

7.

F. Jaouen, "O2 Reduction Mechanism on Non-Noble Metal Catalysts for PEM Fuel Cells.
Part II: A Porous-Electrode Model To Predict the Quantity of H2O2 Detected by Rotating
Ring-Disk Electrode," The Journal of Physical Chemistry C, 113(34), 15433-15443 (2009).

8.

H. S. Wroblowa, P. Yen Chi and G. Razumney, "Electroreduction of oxygen: A new
mechanistic criterion," Journal of Electroanalytical Chemistry and Interfacial
Electrochemistry, 69(2), 195-201 (1976). doi:10.1016/S0022-0728(76)80250-1

9.

K. L. Hsueh, D. T. Chin and S. Srinivasan, "Electrode kinetics of oxygen reduction: A
theoretical and experimental analysis of the rotating ring-disc electrode method," Journal of
Electroanalytical Chemistry and Interfacial Electrochemistry, 153(1), 79-95 (1983).
doi:10.1016/S0022-0728(83)80007-2

10. A. Damjanovic, M. A. Genshaw and J. O. M. Bockris, "The Role of Hydrogen Peroxide in
the Reduction of Oxygen at Platinum Electrodes," The Journal of Physical Chemistry,
70(11), 3761-3762 (1966). doi:10.1021/j100883a515

79

11. A. J. Bard and L. R. Faulkner, Electrochemical methods : fundamentals and
applications,(John Wiley, New York, 2000).
12. T. S. Olson, S. Pylypenko, J. E. Fulghum and P. Atanassov, "Bifunctional Oxygen
Reduction Reaction Mechanism on Non-Platinum Catalysts Derived from Pyrolyzed
Porphyrins," Journal of the Electrochemical Society, 157(1), B54-B63 (2010). doi:Doi
10.1149/1.3248003
13. A. Bonakdarpour, M. Lefevre, R. Z. Yang, F. Jaouen, T. Dahn, J. P. Dodelet and J. R. Dahn,
"Impact of loading in RRDE experiments on Fe-N-C catalysts: Two- or four-electron
oxygen reduction?," Electrochemical and Solid State Letters, 11(6), B105-B108 (2008).
doi:10.1149/1.2904768
14. N. Guillet, L. Roue, S. Marcotte, D. Villers, J. P. Dodelet, N. Chhim and S. Trevin,
"Electrogeneration of hydrogen peroxide in acid medium using pyrolyzed cobalt-based
catalysts: Influence of the cobalt content on the electrode performance," Journal of Applied
Electrochemistry, 36(8), 863-870 (2006).
15. S. Marcotte, D. Villers, N. Guillet, L. Roue and J. P. Dodelet, "Electroreduction of oxygen
on Co-based catalysts: determination of the parameters affecting the two-electron transfer
reaction in an acid medium," Electrochimica Acta, 50(1), 179-188 (2004).
16. N. Leonard, V. Nallathambi and S. C. Barton, "Carbon Supports for Non-Precious Metal
Oxygen Reducing Catalysts," Journal of the Electrochemical Society, 160(8), F788-F792
(2013). doi:10.1149/2.026308jes
17. M. H. Robson, A. Serov, K. Artyushkova and P. Atanassov, "A Mechanistic Study of 4Aminoantipyrine and Iron Derived Non-Platinum Group Metal Catalyst on the Oxygen
Reduction Reaction," Electrochimica Acta (2012).
18. V. Nallathambi, N. Leonard, R. Kothandaraman and S. Calabrese Barton, "Nitrogen
Precursor Effects in Iron-Nitrogen-Carbon Oxygen Reduction Catalysts," Electrochemical
and Solid State Letters, 14(6), B55-B58 (2011). doi:10.1149/1.3566065
19. U. A. Paulus, T. J. Schmidt, H. A. Gasteiger and R. J. Behm, "Oxygen reduction on a highsurface area Pt/Vulcan carbon catalyst: a thin-film rotating ring-disk electrode study,"
Journal of Electroanalytical Chemistry, 495(2), 134-145 (2001).
20. E. Claude, T. Addou, J. M. Latour and P. Aldebert, "A new method for electrochemical
screening based on the rotating ring disc electrode and its application to oxygen reduction
catalysts," Journal of Applied Electrochemistry, 28(1), 57-64 (1998).
21. K. E. Gubbins and R. D. Walker, "The solubility and diffusivity of oxygen in electrolytic
solutions," Journal of The Electrochemical Society, 112(5), 469-471 (1965).
22. W. M. Haynes (ed.) CRC Handbook of Chemistry and Physics (CRC Press, 2010).

80

23. D. M. H. Kern, "The Polarography and Standard Potential of the Oxygen-Hydrogen
Peroxide Couple," Journal of the American Chemical Society, 76(16), 4208-4214 (1954).
doi:10.1021/ja01645a059
24. H. S. Fogler, Elements of Chemical Reaction Engineering,(PHI Learning Private Limited,
New Dehli, 2009).

81

Chapter 5. Modeling Low-Temperature Fuel Cell Electrodes using Non-precious Metal
Catalysts1
5.1. Abstract
An electrode-scale, transport model for a proton-exchange-membrane fuel cell (PEMFC)
cathode is presented. The model describes the performance of non-precious metal catalysts for
the oxygen reduction reaction in a fuel cell context. Because of its relatively high thickness,
emphasis is placed on phenomena occurring in the cathode layer. Water flooding is studied in
terms of its impact on gas-phase transport and on electrochemically accessible surface area
(ECSA). Although cathode performance in both air and oxygen are susceptible to ECSA loss,
gas diffusion limitations at high current density in air are more significant. In oxygen, catalyst
utilization at high current density is primarily limited by conductivity. For this reason, air fuel
cell data is recommended over oxygen data for characterizing catalyst performance. Due to both
ohmic and mass transport limitations, increased loading of low-cost catalysts does not
necessarily lead to higher performance. Therefore, careful optimization of catalyst layer
thickness is required.

1

Authors: Nathaniel D. Leonard,1* Kateryna Artyushkova,2 Barr Halevi,3 Alexey Serov,2
Plamen Atanassov,2 Scott Calabrese Barton1
1

Department of Chemical Engineering and Materials Science, Michigan State University, East
Lansing, MI 48824(USA)
2
University of New Mexico, Chemical & Nuclear Engineering Department, UNM Center for
Emerging Energy Technologies, University of New Mexico, Albuquerque, NM 87131 (USA)
3
Pajarito Powder, Albuquerque, NM 87102 (USA

82

5.2. Introduction
The high environmental costs of current energy systems drives a search for commercializable
energy technologies with low carbon footprints. Proton exchange membrane fuel cells present
one energy option for transportation applications. A significant impediment to commercialization
has been the cost and availability of catalysts for the oxygen reduction reaction (ORR) due to
prevalent use of platinum group metals (PGM). Metal-Nitrogen-Carbon (MNC) catalysts are a
potential solution to the cost and availability challenges that come with using PGMs. MNC
catalysts are synthesized by mixing metal, carbon, and nitrogen precursors followed by one or
more pyrolysis steps at 700-900 C.1-7 Such MNC catalysts are generally washed in acid to
remove excess metal precursors.1-5 These catalysts generally involve lower volume-specific
activity than platinum, which results in thicker electrodes. It has been previously proposed that a
low-cost catalyst may be allowed to have 10-fold lower activity than platinum, as long as the
catalyst layer was 10-fold thicker.8 This proposal was based on the assumption that transport
loses would not be significant in this thickness regime. The purpose of this paper is to
understand quantitatively how the thickness of these electrodes will impact membrane electrode
assembly (MEA) performance for MNC catalyst.
A number of models have addressed electrode scale transport issues.9-15 These models
generally concentrate on water flooding in the gas diffusion layer (GDL) and consider a
relatively thin catalyst layer. Here, gas and liquid transport in the catalyst layer are considered in
a way that is similar to treatment of gas diffusion layers in previous models. We also treat the
GDL consistently with previous models.
Multiple works have considered the catalyst layer.16-22 These works have primarily
concentrated on thin catalyst layers that do not provide significant gas diffusion limitations in

83

comparison to the GDL. For this reason these works concentrate on how flooding impacts
electrochemically active surface area (ECSA). Here, we consider the impact of flooding on
ECSA in addition to gas diffusion. The ECSA impact is considered by incorporating a flooding
term in the kinetics model similar to Eikerling et al.20 Treatment of gas diffusion limitations are
based on the decrease in effective porosity due to flooding, as well as the dilution of oxygen by
to water vapor and nitrogen, similar to the treatment of the GDL by Weber et al.11 This second
consideration is likely more significant than in thin Pt-based catalyst layers due to increased
catalyst layer thickness as well as the increased hydrophilicity that can accompany heat-treated
carbon materials.
5.3. Experimental
5.3.1

Materials

Catalysts were synthesized by Pajarito Powder using a scaled up proprietary method similar to
the work by Serov et al.23 To summarize, a silica template material is thoroughly mixed with iron
nitrate and various nitrogen precursors. After pyrolyzing the mixture, the silica template is
removed by etching in hydrofluoric acid. A second pyrolysis follows etching.
5.3.2

MEA Fabrication

Catalyst ink containing a 9:11 mass ratio of ionomer to catalyst with 500 mg catalyst per 20
mL solvent was sprayed on a 5 cm2 Sigracet 25BC GDL using a Sono-tek Exacta-Coat
automated spray system. The catalyst layer (CL) plus GDL was assembled with a 211 membrane
and an anode containing 0.2 mgPt/cm2. This assembly was hot pressed at 131°C and 90 psi for six
minutes.

84

5.3.3

Electrochemical Characterization

MEAs, fabricated as explained in previous section, were loaded onto single serpentine pattern
graphite flow plates and assembled with 4.5 N torque. The cells was allowed to equilibrate under
a feed of 200 sccm, 1.5 bar total pressure. After reaching a temperature of 80 °C, they were
broken in by holding for 15 minutes at 0.3 and 0.6 V. MEA activity was measured
potentiostatically at various backpressures over a potential range from open-circuit voltage
(OCV) to 0.3 V, and then back to OCV. At each potential, a current was recorded after a one
minute hold, and a high frequency resistance measurement was recorded. The MEA was poised
at 0.6 V for 15 minutes between polarization measurements. For additional break-in, an initial
polarization from OCV to 0.3 V and back was discarded. Inlet gases were hydrogen and air at
100% relative humidity and 200 sccm.
5.3.4

Physical Characterization

Catalyst layer material for physical characterization was created by spraying the catalyst ink
onto a glass plate. Catalyst ink contained a 9:11 mass ratio of ionomer to catalyst with 500 mg
catalyst per 20 mL solvent. After the material was dry, it was scraped of and used for porosity
measurements.
The pore size distribution (PSD) of catalyst layer material was measured via mercury
intrusion porosimetry (Micromeritics AutoPore IV 9500.) A 3 mL penetrometer with a stem
volume of 1.1 mL was loaded with 0.1 g of CL material. Mercury intrusion was measured over

85

the pressure range of 0.1 to 30000 psia. The pressure is related to a pore radius with the
Washburn equation:24
rp =

2 cos ✓
pmeas

[1]

where the measured pressure, pmeas, contact angle, θ, and surface tension of mercury, γ, establish
the pore radius, rp. In this way the change in mercury volume from one pressure to another can
be associated with intrusion into a certain pore size and a pore size distribution can be calculated.

5.4. Model Description
5.4.1

Overview

This section describes the governing equations for a two-region, multi-phase, transport model
of an electrolyte membrane - catalyst layer (CL) - gas diffusion layer (GDL) assembly. This onedimensional, steady-state model consists of two regions (CL and GDL) with the electrolyte
membrane being considered as a boundary condition at the CL surface (Figure 5.1). Each region
has different properties, but the governing equations are similar.

86

Figure 5.1: Schematic of cathode showing different phases and regions of the model with
boundary conditions (black-outlined, white arrows), transport phenomena (blue arrows), and
generation terms (red arrows).

Within each region there are four phases: an electron-conducting solid phase, an ionconducting ionic phase, and non-conducting liquid and gas phases. The primary dependent
variables in these phases are electronic phase potential, Ve, ionic phase potential, Vi, liquid
pressure, PL; and vapor fractions of oxygen and water vapor, xo and xw. Currents and species flux
are also dependent variables, but can all be related to gradients of primary dependent variables,
as discussed in detail below. The governing equations consider the conservation of electrons,
protons, liquid water, and water vapor (Equations 2, 3, 4, and 5 respectively):

87

0=

e r2 Ve + nF rORR

[2]

0=

i r 2 V i

[3]

0=

nF rORR

kL 2
r PL + 2rORR
µ

revap

rNw = revap

[4]

[5]

where κe is electronic conductivity, Ve is electronic phase potential, n is electrons per mole
oxygen, F is Faraday’s constant, rORR is an oxygen reduction reaction rate, κi is ionic
conductivity, Vi is ionic potential, kL is liquid permeability (a function of saturation), µ is
viscosity, PL is liquid pressure, revap is an evaporation rate, and Nw is flux of water vapor. Flux of
nitrogen in air is set to zero everywhere, because it does not participate in any reaction, and the
membrane is considered to be gas-impermeable.
Boundary conditions for the governing equations fall into three categories: membrane, CLGDL interface, and channel (shown in Figure 5.1 in the black-outlined, white arrows). The
membrane is treated as an electrical insulator (ie =0), with ionic resistance, Rmem, that leads to an
ohmic drop Vi=ii*Rmem in the electrolyte phase. Electroosmotic flux of water through the
membrane is calculated based on flux coefficient, βmem, via the equation: NL= ii*βmem/F.25,26 The
coefficient is a function of pressure as discovered by Jansen and shown in Table 4.26 The
membrane is also considered gas impermeable (Nw= Nn = No =0). At the CL-GDL interface, all
variables are considered continuous. The channel boundary is treated as an ionic insulator (ii =0),
and the solid phase overpotential at the channel is Ve=ie*Rext to account for external resistances
88

not included in this model (current collectors, contact resistances, bipolar plates, anode, and
anode GDL). The resistance Rext was estimated by subtracting GDL, membrane, and CL
resistances from measured high frequency resistance. Using this approach, experimental data did
not have to be iR corrected prior to fitting. At the channel boundary, all concentrations are fixed
to bulk conditions for the inlet gas. The liquid phase pressure is in equilibrium with the gas phase
pressure, making the capillary pressure zero.
5.4.2

Porous Phase

Void space in the catalyst layer consists of water-filled hydrophilic pores with contact angles
less than 90°, and hydrophobic pores with contact angle θobs ≥ 90º, and filled by a combination of
liquid and gas phases. The fraction of pores that are flooded with water is given by saturation, S,
a measure of flooding. Saturation is in turn controlled by the capillary pressure, pc, which is the
pressure difference between the liquid phase and gas phase. Saturation impacts electrochemically
active surface area as well as effective permeability and gas diffusivity. For hydrophilic pores,
S = 1, because capillary pressure is always positive due to the generation of liquid water via
oxygen reduction: the liquid flux results in a pressure gradient away from the channel. This
gradient and boundary condition ensure all liquid pressures are greater than the gas pressure,
which is considered constant and equal to the channel gas pressure.
In order consider hydrophobic pores, a contact angle is required. The contact angle was
measured to be 108 ± 5° based on the average and standard deviation of eight contact angle
measurements. For hydrophobic pores, the capillary pressure, pc, apparent contact angle, θobs,
and surface tension of water, γ, define a critical pore radius, rc, via the Washburn equation:
10,11,27,28

89

rc =

2 cos ✓obs
pc

[6]

The saturation, S, of hydrophobic pores can be calculated by summing the volume of pores
larger than the critical radius, rc. This can be performed via an integration of the pore size
distribution:10

R1

c
S = Rr1

0

dV
dr
dV
dr

dr
dr

[7]

Essentially, as capillary pressure increases, water is forced into increasingly smaller pore sizes.
Although direct integration of a measured pore size distribution (PSD) is achievable, a set of lognormal PSDs (LNPSDs) can be used to describe the PSD to make further calculations more
manageable.11 The use of lognormal PSDs also allows a means to differentiate pore modes. From
the PSDs shown in Figure 5.2 it is evident that there are three different characteristic pore sizes:
a small narrow pore size around 5 nm radius and a broad pore mode at around 300 nm radius (for
neat catalyst) with a shoulder representing another characteristic pore size at a slightly smaller
radius (at 100 nm for neat catalyst). Performing porosimetry of pellets composed of catalyst
layer material (containing ionomer) allows the impact of the ionomer on porosity to be observed.
By comparing porosity with and without ionomer it is evident that ionomer content primarily
impacts macroporous and larger mesopores. The deviation between the two curves occurs around
25 nm pore radius which agrees well with literature values indicating that ionomer agglomerate
size is about 40 nm diameter.29 Furthermore, a large portion of the porosity, ~80%, is situated
initially in macropores, and with the addition of ionomer the overall macroporosity decreases as
well as the size of the macropores. This indicates that the macropores are covered with an
ionomer film reducing the effective pore diameter.
90

Figure 5.2: Pore size distribution of catalyst as well as three different pellets of catalyst layer
material (including ionomer) as measured by mercury intrusion porosimetry with a fit composed
of three lognormal pore size distributions. Fit characteristics are included in Table 5.1.

Table 5.1: Pore Size Distribution Fit Parameters
Parameter
Mean
Radius
/nm
Spread
Volume
fraction

Pore
Mode
1
2
3
1
2
3
1
2
3

Neat Catalyst
4.57 ± 0.04
164.6 ± 8.3
307 ± 8
0.222 ± 0.009
1.115 ± 0.035
0.320 ± 0.032
0.191 ± 0.033
0.656 ± 0.023
0.153 ± 0.022

Pellet
(0.222 gcat/cc)
4.11 ± 0.03
98.6 ± 8.8
121.8 ± 2.4
0.222 ± 0.008
1.097 ± 0.080
0.293 ± 0.026
0.266 ± 0.045
0.524 ± 0.033
0.210 ± 0.030

91

Pellet
(0.304 gcat/cc)
4.08 ± 0.03
84.4 ± 17.5
113.3 ± 3.8
0.225 ± 0.008
1.242 ± 0.140
0.416 ± 0.047
0.263 ± 0.093
0.451 ± 0.069
0.286 ± 0.063

Pellet
(0.336 gcat/cc)
4.06 ± 0.02
51.9 ± 47.0
102.0 ± 3.5
0.214 ± 0.007
1.943 ± 0.623
0.586 ± 0.046
0.261 ± 0.144
0.306 ± 0.109
0.433 ± 0.094

The various distributions are easily modeled by considering the porosity to be composed of
three LNPSDs as shown in Figure 5.2. Each LNPSD, Vk(r), can be represented as a function of a
characteristic pore radius, rk, spread, sk, and fraction of total porosity, fk:10

Vk (r) = fk

"

1
p

rsk 2⇡

exp

✓

1 ln r ln rk
p
2 sk 2

◆2 #

[8]
Saturation is calculated by integrating these PSDs over the whole range for the hydrophilic pore
fraction (because pc >0) and for all pores greater than the critical radius for the hydrophobic pore
fraction:10
Sk = fHI,k + fHO,k [1

erf (

ln(rc ) ln(rk )
p
)]
sk 2

[9]

where the saturation of the kth LNPSD, Sk, is a function of the hydrophilic pore fraction, fHI, and
the hydrophobic pore fraction, fHO.
Saturation impacts transport parameters via the permeability, kL, and a reduction in effective
diffusivity. In the liquid phase, an expression used by Weber et al accounts for tortuosity as well
as weighting of pore sizes according to Poiseuille flow.10,11

kL = ksat

S2
fH [1
2

erf (

ln(rc ) ln(rk )
p
sk 2

p
sk 2)]

[10]

In the gas phase, the governing mass balance on the various constituents is coupled with
Stefan-Maxwell multicomponent diffusion: 10,21
rxi =

X Ni xj
i6=j

92

Nj x i

cDij

[11]

where xi represents mole fraction, Ni is the flux, c the total molar concentration, and Dij a
binary diffusion coefficient. The diffusion coefficient is an effective value due to the impact of
porosity on the mean free path of constituents:22,27,28
0
Dij,ef f = Dij
(✏(1

S))3/2

[12]

where the bulk diffusivity, !!"! , is modified by the porosity (ε) and saturation (S from Eq.9). The
effective diffusivity is also modified to account for Knudsen diffusion. Knudsen diffusion limits
diffusivity in small pores where wall effects dominate. In order to incorporate the impact of
Knudsen diffusion, a mean open pore radius is calculated. This mean open pore radius, rK is used
to calculation Knudsen diffusivity, DK:11,28
2rK
DK =
3

✓

8RT
⇡M

◆1/2

[13]

where R is the ideal gas constant, T is the temperature, and M is the molar mass of the gas
molecule. Contributions of the two diffusion modes can be combined via the Bosanquet
equation:30-33
1
1
1
=
+
D
DK
DM

[14]

where the diffusivity, D, is a function of Knudsen diffusivity, DK, and molecular diffusivity, DM.
Darcy’s Law governs liquid phase transport where liquid flux is proportional to the liquid
pressure gradient: 11,28
kL
rPL
µ

NL =

93

[15]

where NL is the flux of liquid water, kL is the effective permeability (a function of saturation, Eq.
10), µ is the viscosity of water, and ∇PL is the liquid pressure gradient.
5.4.3

Conductive Phases

Two different conductive phases exist: solid electronic (s) and electrolyte (e) conductor. Here,
Ohm’s Law governs charge transport:21
ii = -κi łVi ; ie = -κe łVe

[16]

where, for a given phase (i or e), the current density, i, is a function conductivity, κ, and
potential, V.
Electronic conductivity was calculated from correlating high frequency resistance (HFR) with
the thickness of MEAs of varying loading. This correlation, shown in Figure 5.3, assumes that
ionic conductivity in the catalyst layer is much lower than electronic conductivity. Essentially, at
high frequencies capacitors behave like short circuits eliminating the effect of transport and
kinetic effects on the impedance. Without these effects the HFR is the ohmic resistance of the
system. In the catalyst layer, the solid and electrolyte systems are in parallel making the HFR
(RHFR):
RHFR = Rsys +

Ri Re
R i + Re

[17]

where the HFR is function of the resistances of the electronic (Re) and ionic phases (Ri) as well
as the resistance of the remainder of the system (Rsys). When Ri >> Re, the right hand side of Eq.
17 simplifies to RHFR = Rsys + Re. By comparing HFR to the electrolyte resistance calculated in
the next section we can confirm this assumption that Ri >> Re. The slope of the HFR with respect
to catalyst layer thickness represents the electric conductivity of the catalyst layer. From this
94

correlation two model parameters can be found: the electronic conductivity, κe= 0.62 ± 0.38
S/cm and the resistance of the remainder of the system (contact resistance, etc.) Rsys= 0.046 ±
0.010 Ω-cm2. While the error on the conductivity is quite large, we consider this a reasonable
estimate for our purposes.

Figure 5.3: Conductivity of catalyst layer material by correlating high frequency resistance (HFR)
with thickness of catalyst layers with varying loadings at constant density.

Electrolyte conductivity was calculated from electrochemical impedance spectroscopy.
Makharia et al34 and Eikerling et al17 show that impedance of catalyst layers can be well modeled
by transmission line circuits. Specifically, when the electrode potential is small and the
electronic resistance in the catalyst layer is insignificant, the impedance, Z, can be modeled by
the expression: 11,17,34

Z = Rext +

p
p
Ri ZCT coth( Ri /ZCT )

95

[18]

where Rext is the resistance external to the catalyst layer, ω is frequency, Ri is the ionic resistance
in the catalyst layer, and ZCT is a charge transfer impedance. This charge transfer impedance is
ostensibly a Voigt element, where the impedance is a function of some charge transfer resistance,
RCT, and a capacitance, CCT:
ZCT =

RCT
1 + j!RCT CCT

[19]

Figure 5.4 shows impedance spectra of two loadings at two different current densities that are all
well below the onset of mass transport limitations; lines represent the fit by the transmission line
model, Eq. 18. From the fitted ionic resistance, Ri, conductivity for the electrolyte in the catalyst
layer was calculated to be κi= 0.0181 ± 0.0049 S/cm. A comparison of this ionic conductivity
with the electronic conductivity (κe= 0.62 ± 0.38 S/cm) calculated in the previous paragraph
confirms the assumption that κi << κe (or Ri >> Re).

96

Figure 5.4: Impedance spectra of MEA with two different loadings at two different current
densities. Fits shown are based on Eq. 18. Fitted parameters are shown in Table 5.2.

Table 5.2: Impedance Fit Parameters
Charge
Electrolyte
Loading / mg Current / A
Transfer
Resistance (Re)
cm-2
cm-2
Resistance
/ Ω cm2
(RCT) / Ω cm2
2
0.01
0.434 ± 0.066 0.468 ± 0.029
2
0.05
0.466 ± 0.074 0.447 ± 0.029
4
0.01
0.580 ± 0.165 0.385 ± 0.046
4
0.05
0.544 ± 0.166 0.322 ± 0.044

5.4.4

External
Capacitance
-2 Resistance (Rext) /
(CCT) / mF cm
Ω cm2
126 ± 18
131 ± 20
151 ± 44
153 ± 48

0.045 ± 0.003
0.044 ± 0.003
0.070 ± 0.005
0.068 ± 0.005

Generation Terms

The generation terms for oxygen reduction and evaporation are treated as first order reaction
rates. Oxygen reduction is expressed in a symmetrical Butler-Volmer form, in which the forward
reaction (oxygen reduction) is first order with respect to the local partial pressure of oxygen:28,35

97

rORR


✓
◆
↵F ⌘
i0,e↵
po exp
=
nF
RT

⌘ = U0

V0

exp

Vi

✓

(1

↵)F ⌘
RT

◆

Ve

[20]

[21]

where po is the partial pressure of oxygen, η is the over-potential, and α the transfer coefficient.
The over-potential, η, is a function of U0, the reversible potential; V0, the electrode polarization
(an independent variable); Vs, the ionic phase potential; and Ve, the electronic potential. The
effective pre-exponential exchange current density, i0,eff, is considered to be a function of
porosity, saturation, and ionomer volume fraction due to the requirement that protons, electrons,
and oxygen have access to catalytic active sites. To incorporate the impact of these phases, the
exchange current density is modeled as a three-phase effective parameter:27

i0,e↵ = i0 fi fs ✏(1

S)

[22]

where fi is the volume fraction of ionomer and represents the probability that protons are
available at any given differential volume element. The ionomer volume is calculated based on
the ionomer content and density. The volume fraction of solid catalyst, fs, represents the
probability that a given volume element has connectivity with the electronic phase. The solid
catalyst volume is calculated by subtracting porosity as measured by porosimetry and ionomer
volume from the volume of the pellets used for porosimetry experiments. The porosity, ε, is
calculated for a given catalyst layer specific volume by subtracting solid catalyst volume and
ionomer volume. The expression ε (1-S) represents the volume fraction of oxygen-accessible
open pores. In the expression above, fsε can also be interpreted as an area per unit volume that is
dependent on the quantity of porous and solid phases. Because the total volume fractions must

98

equal one (1= fs + fi + ε), an increase in one of these volume fractions necessitates a decrease in
the remaining volume fraction.
The evaporation term, revap, is dependent on water vapor pressure and the partial pressure of
water: 11,28

revap = km aT (Pvap

pw )

[23]

where km is a mass-transfer coefficient, aT the area per unit volume, Pvap, the vapor pressure of
water, and pw the partial pressure of water vapor. Because the reaction is fast and at
equilibrium,11,36 the rate can be thought of as the necessary evaporation necessary to keep the
vapor pressure and water partial pressure equal.
5.4.5

Model Implementation

The model was written with MATLAB using the boundary value solver BVP5C. The
boundary value solver generated the mesh density necessary to achieve 0.1% error for residuals
above the threshold value 10-6. Generally, the number of points was on the order to 20 in lowcurrent regimes and 200 in high current regimes. A nonlinear, least-squares fitting function was
used to fit the following parameters: hydrophobic pore fraction, fHO; saturated liquid
permeability, kw; exchange current density, i0; reversible potential, U; and exchange coefficient,
α. The parameters were fitted simultaneously to polarization data at four different air pressures
and nine potentials in the range 0.82 to 0.30 volts. Parameters were fit to a tolerance of 0.1%.
The current density was calculated as the electronic current at the channel.
Sensitivity was calculated by running the model at a fixed step size of 1% for a given
parameter. The change in current density, ∆i, due to a change, ∆x, in the value of a parameter x,
leads to the sensitivity, Bx:
99

Bx =

i/i
x/x

[24]

A positive sensitivity indicates a correlation between current and the parameter of interest,
whereas a negative value indicates anti-correlation. No correlation is indicated by a sensitivity of
zero. The model was non-dimensionalized as shown in Appendix A.

5.5. Results and Discussion
The model was used to fit MEA data obtained at three different backpressures: 2, 12, and 30
psig. These pressures correspond to partial air pressures: 0.5, 1.2, and 2.4 atm respectively
correcting for altitude in Albuquerque, NM (Patm=0.83 bar)37 and experimental conditions.
Catalyst loadings of 2, 3, and 4 mg/cm2 were studied. Five parameters were fit: hydrophobic pore
fraction, fHO; saturated liquid permeability, kw; exchange current density, i0; reversible potential,
U0 and transfer coefficient, α. The fitted parameters are shown in Table III, while the remaining
parameters are shown in Table IV at the end of the chapter.
A plot of experimental polarization curves along with the fitted model is shown in Figure 5.5.
From observation, the model fits the data well with regards to the shape of the curves. The model
fit shown in Figure 5.5 can be further validated by comparing fitted values with expected values.
The exchange coefficient, α is 0.407 ± 0.009 which is in the range of typical values (0.3-0.7).35
The exchange current density is large because it represents a theoretical value if all points in the
electrode where on a triple-phase boundary. This is not possible for materials of finite thickness,
so it is best to consider the effective exchange coefficient (Eq. 22) for comparison purposes. The
effective exchange current densities are 28, 42, and 56 mA/cm2 at 30 psi for 2, 3, and 4 mg/cm2
loadings respectively. These numbers correspond well with the currents at which the
100

experimental polarization curves deviate from the Tafel regime in the upper part of Figure 5.5.
At some point, when the active site for these catalysts is better understood, i0 should be revisited
to account for turnover rate and adsorption effects. The fitted permeability of (5.27 ± 2.58)×10-13
agrees well with the value predicted by the Kozeny-Carmen Equation (4×10-13).12 The
hydrophobic pore fraction was found to be 0.287 ± 0.110. Based on the assumption that all
hydrophobicity is due to the ionomer, this number can be compared to the electrolyte solid
fraction in the catalyst layer, 0.193 ± 0.070. The agreement between these values corroborates
the theory that the majority of the hydrophobicity is due to the ionomer.

Figure 5.5: Comparison of experimental and model results at three different gas pressures for (a) 2
2
2
2
2
mg/cm , (b) 3 mg/cm , and (c) 4 mg/cm . Conditions: Temp. 80 C, 25BC GDL, 4mg/cm Catalyst
2
Loading, 45 wt.% Nafion, Cell Area: 5cm , membrane: NR211.

Table 5.3: Fitted Parameters
Parameter
Transfer Coefficient
Exchange current density (i0)
Reversible potential (U0)
CL Hydrophobic pore fraction (fHO)
Saturated Liquid Permeability (kw)

Values
0.416 ± 0.010
1300 ± 512
0.853 ± 0.005
0.287 ± 0.110
(5.03± 2.41)×10-13

Units
A / bar cm3
V
cm2

Now that the fit has been established, profiles in the catalyst layer can be used to better
understand transport phenomena. Figure 5.6 shows electronic current profiles (a), oxygen vapor
fraction profiles (b), and flooding (c) in the catalyst layer at three different potentials. How these

101

potentials correspond to the polarization curve is shown in Figure 5.6d: points I, II, and III
correspond to potentials in the kinetically limited regime, ohmic limited regime, and catalyst
layer oxygen transport limited regime respectively. The first point shows a linear current profile
as shown in Figure 5.6a, which is typical of kinetically-limited ORR in the absence of
concentration gradients. Point II shows an ohmically-limited current profile, in which ohmic
resistance in the catalyst layer electrolyte cause most of the ORR to occur near the membraneCL interface, where a steep current gradient is observed. Continuing to lower potentials, in
Figure 5.6b, point III shows conditions where the catalyst layer closest to the membrane has been
completely exhausted of oxygen. Oxygen depletion makes those sections inert, and the current
profile is flat and zero as shown in Figure 5.6a between 0 and 0.1. The limiting case (not shown)
is where the catalyst layer is almost entirely oxygen starved and all the ORR occurs at the CLGDL interface. At this extreme, the polarization becomes linear due to the ionic resistance of the
catalyst layer.

Figure 5.6: (a) Solid phase current distribution, (b) ionic over-potential, and (c) flooding at
dimensionless positions in the catalyst layer at three different potentials on the polarization curve
(d) showing I, kinetic limitations; II, electrolyte conductivity limitations; and III, oxygen diffusion
limitations. For position in the catalyst layer, zero represents the membrane interface and one
represents GDL interface.

102

A sensitivity study of performance in air is shown in Figure 5.7. Sensitivity is defined as a
ratio of relative change in performance (current) to relative change in a given parameter. Figure
5.7 shows sensitivity as it varies with potential (on the y-axis). At all potentials hydrophobic pore
volume shows high sensitivity, this indicates that flooding is high. Essentially, flooding limits
electrochemically active surface area at high potentials (kinetically limited, I) and limits oxygen
diffusion at low potentials (oxygen transport limited, III). Around 0.6 V there is a dip in the
hydrophobic pore fraction sensitivity. This dip correlates to an increase in electrolyte
conductivity sensitivity. This confirms the ohmic limited regions identified by point II in Figure
5.6. The significant anti-correlation is loading: at high potentials increased loading means
increased electrochemically active surface area, but at low potentials this increased loading only
inhibits oxygen diffusion, leading to lowered utilization and reduced performance.

Figure 5.7: Plot of current sensitivity in air varying with potential.

103

A sensitivity study can also be calculated for oxygen, as shown in Figure 5.8. The sensitivity
in oxygen is similar to air at high potentials, but not at low potentials. The similarity in
sensitivity is expected at high potentials due to kinetic limitation. Once outside of this regime,
the similarities decrease substantially due to the lack of transport limitations in oxygen. In this
ohmically-limited regime flooding only impacts electrochemically active surface area causing
the hydrophobic pore fraction sensitivity to decrease as kinetics become less limiting at lower
potentials. In oxygen, electron transport limits current density at low potentials. This is shown by
the increase in ionic conductivity sensitivity around 0.8 V. At low potentials activity reaches a
mixed domination region. In this region performance is equally controlled by hydrophobic
porosity, ionic conductivity, and kinetic parameters. Essentially the current profile is similar to
point II in Figure 5.6a, but as the potential decreases further the current distribution becomes
increasingly concentrated at the CL-membrane interface due to ionic conductivity limitations. As
the current distribution becomes more concentrated at that interface, less charge is conducted
through the catalyst layer electrolyte, which decreases the ionic conductivity sensitivity. External
ohmic losses also become more significant at low potentials. This value is not a fitted parameter,
but it does include contact resistance between the CL and GDL due to the high Nafion content in
the CL. For this reason, external resistance was added to the oxygen sensitivity study (purple line
with open diamonds) and represents the impact of contact resistance between assembly layers.

104

Figure 5.8: Plot of current sensitivity in oxygen varying with potential.

To return to air sensitivity, two recommendations can be made for future non-PGM catalysts
and their application in fuel cells. Loading should be optimized with current density at key
operating potentials in mind. For applications running at max power, lower loadings are desired
to increase performance in the 0.6-0.4 V range. A second observation that can be gleaned from
the air sensitivity is that increasing ionomer content near the CL-membrane interface could
mitigate ohmic losses resulting from poor catalyst utilization at low potentials, as previously
observed by other workers for platinum based catalyst layers.38,39 This would be the result of
optimizing catalyst ink for different depths in the catalyst layer, and the catalyst layer close to the
GDL would be optimized for oxygen diffusion, while the catalyst layer close to the membrane
would be optimized for electrolyte conductivity.

105

Figure 5.9: (a) Optimal loading as it varies with potential, (b) current as it varies with loading at
various potentials, (c) current as it varies with density and loading at 0.8 V, and (d) current as it
varies with density and loading at 0.4 V. Conditions: 30 psi, Temp. 80 C, 25BC GDL, 45 wt.%
2
Nafion, Cell Area: 5cm , membrane: NR211.

Finally, optimization studies were conducted for loading and catalyst layer density. For each
parameter, the value was obtained that provided maximum current density with all other
parameters at baseline values (Table 3 and 4). From Figure 5.9a, it is apparent that optimal
loading changes with potential, because mass transport limitations are more significant at low
potential and kinetic limitations more significant at high potential. CL density represents how
compressed the catalyst layer is during MEA construction and the optimal value does not vary
much with potential as shown in contour plots Figure 5.9c and 5.9d.

106

5.6. Conclusions
A one-dimensional, steady-state model was developed for electrode-scale transport in a nonPGM PEMFC cathode. It was found that flooding limits oxygen transport in air due to reduced
effective porosity and active surface area. At low potentials the flooding limitations result in
poor catalyst utilization and increased ohmic losses. In contrast to air performance, oxygen
performance depends primarily on active surface area and electron conductivity. The difference
in sensitivity between air and oxygen performance suggests that oxygen MEA performance is
not an effective indicator for non-PGM catalysts.
Finally, cathodes can be designed for improved air performance by improving catalyst
utilization. Utilization can be improved in two different ways. Most simply, utilization can be
increase by decreasing loading. Although this model predicts improved performance with
decreased loading, it is important to note that as catalyst thickness decreases, the relative
importance of interfacial effects between CL and GDL will increase. Utilization can also be
improved by optimizing sections of the catalyst layer for different phenomena. Optimization
would include improving gas diffusion near the GDL and improving ionic conductivity near the
membrane.
5.7. Acknowledgments
We gratefully acknowledge the partial financial support from the U.S. Department of Energy
(EERE), under a Non PGM Catalyst development effort (Contract no EE 0000459, lead by
Northeastern University, Sanjeev Mukerjee, P.I.).

107

Table 5.4: Model Parameters
Parameter
Diffusion coefficient of
oxygen in water (pDow) 40
Diffusion coefficient of
oxygen in nitrogen (pDon)

Values
0.3022× (T/323.83)2.334

Units
bar cm2 / s

0.0544× (T/143.01)1.823

bar cm / s

Literature
Value11,40,41

Diffusion coefficient of
water in nitrogen (pDwn) 40
Evaporation rate (kmaT) 11

0.2599× (T/299.42)2.334

bar cm2 / s

100

Viscosity (µ) 11

(2695.3-6.6 T)×10-11

mol / bar s
cm3
bar s

Literature
Value11,40,41
Literature Value11,36

Water surface tension (γ)

0.12398-0.00017393×T

N/m

Literature Value11

CL electronic conductivity
(κe)
CL ionic conductivity (κi)

0.62 ± 0.38

S/cm

0.0158 ± 0.0020

S/cm

Catalyst density
CL Hydrophobic Contact
Angle ( obs)
Saturated Permeability
(ksat)12
External Resistance (Rext)

0.32
90.02

gcat/cc
Degrees

Correlated from
HFR
Calculated from
impedance model
Measured
Literature Value11

1×10-9

cm2

Literature Value11

0.046 ± 0.010

Ω cm2

Electro-osmotic coefficient

0.5×Pvap/(Pg-Pvap)

Calculated from
High Frequency
Impedance
Literature26

2

40

11

108

Notes
Literature Value11,
33, 34

Literature Value11

APPENDIX

109

Appendix A Non-Dimensional Cathode Model
This section describes the non-dimensional equations for a multi-phase, transport model of a
catalyst layer (CL). This one-dimensional, steady-state model consists of four phases: an
electron-conducting solid phase, an ion-conducting ionic phase, and non-conducting liquid and
gas phases. The primary dependent variables in these phases are solid phase potential, Vs,
electrolyte phase potential, Ve, liquid pressure, PL; and vapor fractions of oxygen and water
vapor, xo and xw. Currents and species flux are also dependent variables, but can all be related to
gradients of primary dependent variables. A more complete discussion of the model is available
in Chapter 5 of this dissertation.
The governing equations consider the conservation of electrons, protons, liquid water, and
water vapor (Equations 2, 3, 4, and 5 respectively):

0=

s r2 Vs + nF rORR

[25]

0=

e r 2 V e

[26]

0=

nF rORR

kL 2
r PL + 2rORR
µ

revap

rNw = revap

[27]

[28]

where κs is solid conductivity, Vs is solid phase potential, n is electrons per mole oxygen, F is
Faraday’s constant, rORR is an oxygen reduction reaction rate, κe is electrolyte conductivity, Ve is

110

electrolyte potential, kL is liquid permeability (a function of saturation), µ is viscosity, PL is
liquid pressure, revap is an evaporation rate, and Nw is flux of water vapor. Flux of nitrogen in air
is set to zero everywhere, because it does not participate in any reaction, and the membrane is
considered to be gas-impermeable.
Non-dimentionalization was based on the gas phase because understanding oxygen transport
is a key goal for this project. In the gas phase, the governing mass balance on the various
constituents is coupled with Stefan-Maxwell multicomponent diffusion: 11,22
rxi = RT

X Ni x j
i6=j

Nj x i

pDij

[29]

where xi represents mole fraction, Ni is the flux, c the total molar concentration, and Dij a
binary diffusion coefficient. This equation is non-dimensionalized by multiplying through with a
length scale and results in:
rxi =

X

Qij xj

Qji xi

[30]

i6=j

Qij =

RT Ni L
pDij

[31]

where Q is a dimensionless flus. Although it would be fine to leave the equations in this system,
it is better keep a consistent frame of reference for the various fluxes. This is accomplished using
oxygen in dry air as a reference system and results in:
rxi =

X

⇧ij (Qi xj

Qj xi )

[32]

i6=j

111

⇧ij =

pDon
pDij

[33]

where Πij is a dimensionless flux resistance of constituent i in j, and Q is a dimensionless flux
that can be related to the dimensional flux:
Ni = Q0 Qi

Q0 =

pDon
RT L

[34]

[35]

where Q0 is the reference flux resulting from the non-dimensionalization. The above analysis has
an analogy across all phases and results in all fluxes having similar magnitudes, which improves
performance of the boundary value solver.
Darcy’s Law can be similarly non-dimensionalized. This equation governs liquid phase
transport where liquid flux is proportional to the liquid pressure gradient: 12,28
kL
rPL
µ

NL =

[36]

where NL is the flux of liquid water, kL is the effective permeability (a function of saturation, Eq.
9 in Chapter 5), µ is the viscosity of water, and ∇PL is the liquid pressure gradient. This can be
non-dimensionalized by a Q-flux:
r

QL =

⇧L =

L /⇧L

µL pDon
RT kL P

where ΠL is a dimensionless liquid flux resistance and ΨL is a dimensionless liquid pressure.

112

[37]

[38]

Two different conductive phases exist: solid electronic (s) and electrolyte (e) conductor. Here,
Ohm’s Law governs charge transport:22
ii = -κi łVi ; ie = -κe łVe

[39]

where, for a given phase (i or e), the current density, i, is a function conductivity, κ, and
potential, V. In order non-dimensionalize these phases, an analogy can be drawn between the
pressure/liquid-flux system of the previous section of the potential/current system. The current
can be thought of as a molar flux, and the potential is non-dimensionalized by the Tafel slope:
Qe =

r

⇧e =

e /⇧e

pDon n
e

[40]

[41]

where Ψe is a dimensionless electronic potential (VeF/RT) and Πe is a dimensionless electronic
resistance that is a function of electrons per mole oxygen, n, and the electronic conductivity, κe.
Qe can be thought of a dimensionless molar flux of oxygen or a current density:
No =

ie
= Q0 Qo = Q0 Qe
nF

[42]

now that the transport equations have been dimensionalized the generation terms will be
considered.
The generation terms for oxygen reduction and evaporation are treated as first order reaction
rates. Oxygen reduction is expressed in a symmetrical Butler-Volmer form, in which the forward
reaction (oxygen reduction) is first order with respect to the local partial pressure of oxygen:28,31

113

rORR


✓
◆
↵F ⌘
i0,e↵
po exp
=
nF
RT

⌘ = U0

pref exp

V0

Vs

✓

(1

↵)F ⌘
RT

◆

[43]

Ve

[44]

where po is the partial pressure of oxygen, pref is a reference pressure chosen to be 1 atm because
the reverse reaction does not contain constituents in the gas phase, η is the over-potential, and α
the transfer coefficient. The over-potential, η, is a function of U0, the reversible potential; V0, the
electrode polarization (an independent variable); Vs, the solid phase potential; and Ve, the
electrolyte potential. This equation is non-dimensionalized based on a local mass balance of
oxygen:
rNo = rORR


✓
◆
↵F ⌘
i0,e↵
po exp
=
nF
RT

pref exp

✓

(1

↵)F ⌘
RT

◆

[45]

by substituting for a dimensionless oxygen flux, and normalizing our length scale the equation
can be put in terms of a reduced Thiele modulus:
rQo =

2
Th

[xo exp (↵ )

2
Th

=

=

xref exp ( (1

P io,ef f RT L2
nF pDo,n

0

e

i

↵) )]

[46]

[47]

[48]

where ΦTh is a Thiele modulus for oxygen reduction, and Ψ is a dimensionless overpotential that
takes into account electrode polarization (Ψ0) and ohmic losses due to electronic (Ψe) and ionic
(Ψi) conductivities.

114

The evaporation term, revap, is dependent on water vapor pressure and the partial pressure of
water: 12,28

revap = km aT (Pvap

pw )

[49]

where km is a mass-transfer coefficient, aT the area per unit volume, Pvap, the vapor pressure of
water, and pw the partial pressure of water vapor. Because the reaction is fast and at
equilibrium,12,32 the rate can be thought of as the necessary evaporation necessary to keep the
vapor pressure and water partial pressure equal. This equation is similarly non-dimensionalized
based on a local mass balance of water vapor:
dNg,w
= revap |x
dx

[50]

by substituting for a dimensionless oxygen flux, and normalizing our length scale the equation
can be put in terms of a reduced Thiele modulus:
dQw
=
d⇠

2
evap

2
evap (Pvap /Pg

=

xw )

km aT RT LPg
pDo,n

where ΠL is a dimensionless liquid flux resistance and ΨL is a dimensionless liquid. Files
associated with the model have been published on gitlab.msu.edu:
https://gitlab.msu.edu/leona148/CathodeModel

115

[51]

[52]

REFERENCES

116

REFERENCES

1.

M. H. Robson, A. Serov, K. Artyushkova and P. Atanassov, "A Mechanistic Study of 4Aminoantipyrine and Iron Derived Non-Platinum Group Metal Catalyst on the Oxygen
Reduction Reaction," Electrochimica Acta (2012).

2.

R. Bashyam and P. Zelenay, "A class of non-precious metal composite catalysts for fuel
cells," Nature, 443(7107), 63-66 (2006). doi:10.1038/nature05118

3.

H. T. Chung, C. M. Johnston, K. Artyushkova, M. Ferrandon, D. J. Myers and P. Zelenay,
"Cyanamide-derived non-precious metal catalyst for oxygen reduction," Electrochemistry
Communications, 12(12), 1792-1795 (2010). doi:10.1016/j.elecom.2010.10.027

4.

F. Jaouen, E. Proietti, M. Lefèvre, R. Chenitz, J.-P. Dodelet, G. Wu, H. T. Chung, C. M.
Johnston and P. Zelenay, "Recent advances in non-precious metal catalysis for oxygenreduction reaction in polymer electrolyte fuel cells," Energy & Environmental Science, 4(1),
114 (2011). doi:10.1039/c0ee00011f

5. F. Jaouen, J. Herranz, M. Lefevre, J.-P. Dodelet, U. I. Kramm, I. Herrmann, P. Bogdanoff, J.
Maruyama, T. Nagaoka, A. Garsuch, J. R. Dahn, T. Olson, S. Pylypenko, P. Atanassov and
E. A. Ustinov, "Cross-Laboratory Experimental Study of Non-Noble-Metal Electrocatalysts
for the Oxygen Reduction Reaction," ACS Applied Materials & Interfaces, 1(8), 1623-1639
(2009). doi:10.1021/am900219g
6.

N. Leonard, V. Nallathambi and S. C. Barton, "Carbon Supports for Non-Precious Metal
Oxygen Reducing Catalysts," Journal of the Electrochemical Society, 160(8), F788-F792
(2013). doi:10.1149/2.026308jes

7.

V. Nallathambi, N. Leonard, R. Kothandaraman and S. C. Barton, "Nitrogen Precursor
Effects in Iron-Nitrogen-Carbon Oxygen Reduction Catalysts," Electrochemical and SolidState Letters, 14(6), B55-B58 (2011). doi:Doi 10.1149/1.3566065

8.

H. A. Gasteiger, S. S. Kocha, B. Sompalli and F. T. Wagner, "Activity benchmarks and
requirements for Pt, Pt-alloy, and non-Pt oxygen reduction catalysts for PEMFCs," Applied
Catalysis B: Environmental, 56(1-2), 9 (2005).

9.

T. E. Springer, T. A. Zawodzinski and S. Gottesfeld, "Polymer Electrolyte Fuel-Cell
Model," Journal of the Electrochemical Society, 138(8), 2334-2342 (1991). doi:Doi
10.1149/1.2085971

10. A. Z. Weber and J. Newman, "Modeling transport in polymer-electrolyte fuel cells," Chem
Rev, 104(10), 4679-726 (2004).
11. A. Z. Weber and J. Newman, "Modeling Transport in Polymer-Electrolyte Fuel Cells,"
Chemical Reviews, 104(10), 4679-4726 (2004).

117

12. U. R. Salomov, E. Chiavazzo and P. Asinari, "Pore-scale modeling of fluid flow through gas
diffusion and catalyst layers for high temperature proton exchange membrane (HT-PEM)
fuel cells," Computers & Mathematics with Applications, 67(2), 393-411 (2014). doi:Doi
10.1016/J.Camwa.2013.08.006
13. J. T. Gostick, M. W. Fowler, M. A. Ioannidis, M. D. Pritzker, Y. M. Volfkovich and A.
Sakars, "Capillary pressure and hydrophilic porosity in gas diffusion layers for polymer
electrolyte fuel cells," Journal of Power Sources, 156(2), 375-387 (2006). doi:Doi
10.1016/J.Jpowsour.2005.05.086
14. J. T. Gostick, M. A. Ioannidis, M. W. Fowler and M. D. Pritzker, "Pore network modeling
of fibrous gas diffusion layers for polymer electrolyte membrane fuel cells," Journal of
Power Sources, 173(1), 277-290 (2007). doi:Doi 10.1016/J.Jpowsour.2007.04.059
15. Q. Ye and T. Van Nguyen, "Three-dimensional simulation of liquid water distribution in a
PEMFC with experimentally measured capillary functions," Journal of the Electrochemical
Society, 154(12), B1242-B1251 (2007). doi:10.1149/1.2783775
16. A. Z. Weber, "Macroscopic Modeling of the Proton-Exchange-Membrane Fuel-Cell
Catalyst Layer," ECS Transactions, 45(2), 71-83 (2012).
17. M. Eikerling and A. A. Kornyshev, "Electrochemical impedance of the cathode catalyst
layer in polymer electrolyte fuel cells," Journal of Electroanalytical Chemistry, 475(2), 107123 (1999). doi:Doi 10.1016/S0022-0728(99)00335-6
18. J. Liu and M. Eikerling, "Model of cathode catalyst layers for polymer electrolyte fuel cells:
The role of porous structure and water accumulation," Electrochimica Acta, 53(13), 44354446 (2008). doi:Doi 10.1016/J.Electacta.2008.01.033
19. M. Eikerling and A. A. Kornyshev, "Modelling the performance of the cathode catalyst
layer of polymer electrolyte fuel cells," Journal of Electroanalytical Chemistry, 453(1-2),
89-106 (1998). doi:Doi 10.1016/S0022-0728(98)00214-9
20. M. Eikerling, "Water management in cathode catalyst layers of PEM fuel cells - A structurebased model," Journal of the Electrochemical Society, 153(3), E58-E70 (2006). doi:Doi
10.1149/1.2160435
21. W. Yoon and A. Z. Weber, "Modeling Low-Platinum-Loading Effects in Fuel-Cell Catalyst
Layers," Journal of The Electrochemical Society, 158(8), B1007-B1018 (2011).
doi:10.1149/1.3597644
22. M. El Hannach, J. Pauchet and M. Prat, "Pore network modeling: Application to multiphase
transport inside the cathode catalyst layer of proton exchange membrane fuel cell,"
Electrochimica Acta, 56(28), 10796-10808 (2011). doi:10.1016/j.electacta.2011.05.060
23. A. Serov, K. Artyushkova and P. Atanassov, "Fe-N-C Oxygen Reduction Fuel Cell Catalyst
Derived from Carbendazim: Synthesis, Structure, and Reactivity," Advanced Energy
Materials, 4(10), n/a-n/a (2014). doi:10.1002/aenm.201301735
118

24. A. W. Adamson and A. P. Gast, Physical chemistry of surfaces, vol. (Wiley, New York,
1997).
25. A. Z. Weber and J. Newman, "Transport in Polymer-Electrolyte Membranes: III. Model
Validation in a Simple Fuel-Cell Model," Journal of The Electrochemical Society, 151(2),
A326-A339 (2004). doi:10.1149/1.1639158
26. G. J. M. Janssen, "A Phenomenological Model of Water Transport in a Proton Exchange
Membrane Fuel Cell," Journal of The Electrochemical Society, 148(12), A1313-A1323
(2001). doi:10.1149/1.1415031
27. P. K. Das, X. G. Li and Z. S. Liu, "Effective transport coefficients in PEM fuel cell catalyst
and gas diffusion layers: Beyond Bruggeman approximation," Applied Energy, 87(9), 27852796 (2010). doi:Doi 10.1016/J.Apenergy.2009.05.006
28. A. Z. Weber, R. L. Borup, R. M. Darling, P. K. Das, T. J. Dursch, W. Gu, D. Harvey, A.
Kusoglu, S. Litster, M. M. Mench, R. Mukundan, J. P. Owejan, J. G. Pharoah, M. Secanell
and I. V. Zenyuk, "A Critical Review of Modeling Transport Phenomena in PolymerElectrolyte Fuel Cells," Journal of The Electrochemical Society, 161(12), F1254-F1299
(2014). doi:10.1149/2.0751412jes
29. M. Uchida, Y. Fukuoka, Y. Sugawara, N. Eda and A. Ohta, "Effects of microstructure of
carbon support in the catalyst layer on the performance of polymer-electrolyte fuel cells,"
Journal of the Electrochemical Society, 143(7), 2245-2252 (1996). doi:Doi
10.1149/1.1836988
30. W. G. Pollard and R. D. Present, "On Gaseous Self-Diffusion in Long Capillary Tubes,"
Physical Review, 73(7), 762-774 (1948). doi:Doi 10.1103/Physrev.73.762
31. J. Becker, C. Wieser, S. Fell and K. Steiner, "A multi-scale approach to material modeling
of fuel cell diffusion media," International Journal of Heat and Mass Transfer, 54(7-8),
1360-1368 (2011). doi:Doi 10.1016/J.Ijheatmasstransfer.2010.12.003
32. T. Hutzenlaub, J. Becker, R. Zengerle and S. Thiele, "Modelling the water distribution
within a hydrophilic and hydrophobic 3D reconstructed cathode catalyst layer of a proton
exchange membrane fuel cell," Journal of Power Sources, 227(0), 260-266 (2013).
doi:10.1016/j.jpowsour.2012.11.065
33. S. Litster, W. K. Epting, E. A. Wargo, S. R. Kalidindi and E. C. Kumbur, "Morphological
Analyses of Polymer Electrolyte Fuel Cell Electrodes with Nano-Scale Computed
Tomography Imaging," Fuel Cells, 13(5), 935-945 (2013). doi:10.1002/fuce.201300008
34. R. Makharia, M. F. Mathias and D. R. Baker, "Measurement of catalyst layer electrolyte
resistance in PEFCs using electrochemical impedance spectroscopy," Journal of the
Electrochemical Society, 152(5), A970-A977 (2005).
35. A. J. Bard and L. R. Faulkner, Electrochemical methods : fundamentals and applications,
vol. (John Wiley, New York, 2001).
119

36. J. H. Nam and M. Kaviany, "Effective diffusivity and water-saturation distribution in singleand two-layer PEMFC diffusion medium," International Journal of Heat and Mass Transfer,
46(24), 4595-4611 (2003). doi:10.1016/S0017-9310(03)00305-3
37. M. L. Salby, International Geophysics, Volume 61 : Fundamentals of Atmospheric Physics,
vol. (Academic Press, Burlington, MA, USA, 1996).
38. Z. Xie, T. Navessin, K. Shi, R. Chow, Q. Wang, D. Song, B. Andreaus, M. Eikerling, Z. Liu
and S. Holdcroft, "Functionally Graded Cathode Catalyst Layers for Polymer Electrolyte
Fuel Cells: II. Experimental Study of the Effect of Nafion Distribution," Journal of The
Electrochemical Society, 152(6), A1171-A1179 (2005). doi:10.1149/1.1904990
39. D. Song, Q. Wang, Z. Liu, M. Eikerling, Z. Xie, T. Navessin and S. Holdcroft, "A method
for optimizing distributions of Nafion and Pt in cathode catalyst layers of PEM fuel cells,"
Electrochimica Acta, 50(16–17), 3347-3358 (2005). doi:10.1016/j.electacta.2004.12.008
40. R. B. Bird, W. E. Stewart and E. N. Lightfoot, Transport phenomena, vol. (J. Wiley, New
York, 2007).
41. T. R. Marrero and E. A. Mason, "Gaseous diffusion coefficients," J. Phys. Chem. Ref. Data,
1(1), 3-118 (1972).

120

!
Chapter 6 Summary of Research Contributions
This dissertation improves performance and understanding of non-precious metal catalysts
(NPMCs) for proton exchange membrane fuel cells. Chapter 2 is an important optimization of
carbon supports for the autogenic pressure metal-nitrogen-carbon catalyst, while Chapters 3 and
4 increase understanding of this catalyst. Chapter 4 also introduces a new method of analyzing
NPMCs. Chapter 5 both increases understanding of NPMC behavior in fuel cells and provides
specific guidance for Pajarito Powder in how best to use their catalyst. This guidance also relates
more generally to improved membrane electrode assembly (MEA) construction for other
NPMCs.
Initially a porosity study of metal-nitrogen-carbon (MNC) oxygen reduction catalysts and
carbon supports found that mesoporosity is critical to high performance. Porous carbon materials
with varying structural and compositional properties were studied for their impact on the
nitrogen content and activity of MNC catalysts prepared using high-pressure pyrolysis. The
carbon materials and resulting catalysts were characterized morphologically, chemically, and
electrochemically. The results indicated that substrates adsorbing the most nitrogen and iron
show the highest activity. Furthermore, a relationship found between mesoporosity and nitrogen
adsorption indicate the importance of transport of precursors to potential active sites. This work
had a role in establishing mesoporosity as one of the key parameters in MNC catalyst
development.1 Currently, some of the highest performing catalysts use pore-formers to ensure a
high level of mesoporosity.2-4
The second project found a correlation between surface alkalinity and catalytic activity for
the autogenic pressure MNC (APMNC). The basic site strength and quantity were calculated by
two different methods, and it was shown that increased Brønsted-Lowry basicity correlates to

!

121!

!
more active catalysts. Future work should elucidate the chemistry found in this MNC catalyst.
Specifically, a correlation of this alkalinity data with RRDE analysis of catalysts with varying
nitrogen content could couple the basicity with a specific role in the oxygen reduction reaction as
discussed in Chapter 4. Another related need for the APMNC is improved understanding of
synthesis reactions, which could be accomplished by analysis of gases produced during
pyrolysis. This could be accomplished with differential thermal analysis (DTA) or differential
scanning calorimetry (DSC).
In Chapter 4, an analysis of peroxide generation using rotating ring-disk electrodes indicated
that ORR proceeds both via a direct four-electron pathway to water at high potentials and an
indirect peroxide pathway at low potentials. The main contribution of this research was increased
understanding of APMNC function. Specifically, oxygen reduction is dominated by the direct
four-electron pathway to water without a desorbing intermediate dominates at higher potentials
because peroxide generation is inhibited due to site availability. At lower potentials, oxygen
reduction begins to shift to the indirect peroxide pathway due to fast kinetics and higher site
availability. The net peroxide generation remains relatively low over the entire range due to
reduction of peroxide to water. This work not only presents data on a significant catalyst in the
literature, but it also represents a novel method for analyzing non-precious metal catalyst via
RRDE.
In order to achieve the full benefit of this work, the steady-state RRDE analysis should be
performed on multiple catalysts for comparison. For example, performing this analysis for
catalysts of differing surface chemistries could isolate the oxygen adsorption properties of
various surface groups. One could also study how the measured adsorption coefficients

!

122!

!
compared to calculated oxygen binding energies. These types of studies could give important
indications of how various catalysts are involved in both reaction pathways.
Chapter 5 presented an electrode-scale, transport model for a proton-exchange-membrane
fuel cell (PEMFC) cathode. The model described the performance of non-precious metal
catalysts in a fuel cell context. Water flooding was studied in terms of its impact on gas-phase
transport and electrochemically accessible surface area (ECSA). Although cathode performance
in both air and oxygen are susceptible to ECSA loss, gas diffusion limitations at high current
density in air are more significant. In oxygen, catalyst utilization at high current density is
primarily limited by conductivity. For this reason, air fuel cell data is recommended over oxygen
data for characterizing catalyst performance. Increased loading of low-cost catalysts does not
lead to higher performance, due to transport limitations. These findings have been useful in
optimization of MEAs for MNC catalysts.
There are two important next steps for this work. First, the model should be validated with a
thin platinum electrode. The purpose of this work would be to confirm a correct distribution of
transport limitations between the catalyst layer and GDL. The second important step is extending
this model to two or three dimensions. The current 1-D model assumes no gradients along the
channel, and does not account for the impact of the land on flooding and transport. These
features are necessary to properly model GDL behavior.
!
!
!
!
!
!
!
!
!

!

123!

!

!

!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
REFERENCES

!

124!

!
REFERENCES

1.

K. N. Wood, R. O'Hayre and S. Pylypenko, Recent progress on nitrogen/carbon structures
designed for use in energy and sustainability applications, Energy & Environ. Sci., 7, 1212
(2014), 10.1039/C3EE44078H.

2.

A. Serov, M. H. Robson, M. Smolnik and P. Atanassov, Templated bi-metallic non-PGM
catalysts for oxygen reduction, Electrochim. Acta, 80, 213 (2012), Doi
10.1016/J.Electacta.2012.07.008.

3.

M. H. Robson, A. Serov, K. Artyushkova and P. Atanassov, A Mechanistic Study of 4Aminoantipyrine and Iron Derived Non-Platinum Group Metal Catalyst on the Oxygen
Reduction Reaction, Electrochim. Acta (2012)

4.

E. Proietti, F. Jaouen, M. Lefevre, N. Larouche, J. Tian, J. Herranz and J. P. Dodelet, Ironbased cathode catalyst with enhanced power density in polymer electrolyte membrane fuel
cells, Nature communications, 2, 416 (2011), 10.1038/ncomms1427.

!

125!